Thesis AnishPandey 512ME119

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/320243391
Mobile Robot Navigation in Static and Dynamic Environments using Various

Soft Computing Techniques
Thesis · July 2016

DOI: 10.13140/RG.2.2.11257.57440
CITATIONS READS
2 1,420
1 author:
Dr. Anish Pandey

KIIT University
71 PUBLICATIONS 1,200 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Wheeled cleaning robot View project
All content following this page was uploaded by Dr. Anish Pandey on 06 October 2017.
The user has requested enhancement of the downloaded file.

Mobile Robot Navigation in Static and
Dynamic Environments using Various
Soft Computing Techniques
Anish Pandey
Department of Mechanical Engineering

National Institute of Technology Rourkela
Mobile Robot Navigation in Static and Dynamic Environments
using Various Soft Computing Techniques
Dissertation submitted to the
in partial fulfillment of the requirements
of the degree of
Doctor of Philosophy
in
Mechanical Engineering
by
Anish Pandey
(Roll Number: 512ME119)
under the supervision of
Prof. Dayal R. Parhi
July, 2016
Department of Mechanical Engineering

i
July 18, 2016
Certificate of Examination
Roll Number: 512ME119
Name: Anish Pandey
Title of Dissertation: Mobile Robot Navigation in Static and Dynamic Environments

using Various Soft Computing Techniques
We the below signed, after checking the dissertation mentioned above and the official
record book (s) of the student, hereby state our approval of the dissertation submitted in
partial fulfillment of the requirement of the degree of Doctor of Philosophy in
Mechanical Engineering at National Institute of Technology, Rourkela. We are satisfied
with the volume, quality, correctness, and originality of the work.
Dayal R. Parhi
Principal Supervisor
Bidyadhar Subudhi Tarapada Roy

Member (DSC) Member (DSC)
Sarat Kumar Das Santosha Kumar Dwivedy

Member (DSC) Examiner
Kalipada Maity
Chairman (DSC)
ii
Dr. Dayal R. Parhi

Professor
July 18, 2016
Supervisor’s Certificate
This is to certify that the work presented in this dissertation entitled “Mobile Robot
Navigation in Static and Dynamic Environments using Various Soft Computing
Techniques” by “Anish Pandey”, Roll Number: 512ME119, is a record of original
research carried out by him under my supervision and guidance in partial fulfillment of
the requirements of the degree of Doctor of philosophy in Mechanical Engineering.
Neither this dissertation nor any part of it has been submitted for any degree or diploma
to any institute or university in India or abroad.
Dayal R. Parhi
iii
To my Parents,
with all my love
iv
Declaration of Originality
I, Anish Pandey, Roll Number: 512ME119 hereby declare that this dissertation entitled
“Mobile Robot Navigation in Static and Dynamic Environments using Various Soft
Computing Techniques” represents my original work carried out as a doctoral student of
NIT Rourkela and, to the best of my knowledge, it contains no material previously
published or written by another person, nor any material presented for the award of any
degree or diploma of NIT Rourkela or any other institution. Any contribution made to
this research by others, with whom I have worked at NIT Rourkela or elsewhere, is
explicitly acknowledged in the dissertation. Works of other authors cited in this
dissertation have been duly acknowledged under the section “Bibliography”. I have also
submitted my original research records to the scrutiny committee for evaluation of my
dissertation.
I am fully aware that in case of my non-compliance detected in the future, the Senate
of NIT Rourkela may withdraw the degree awarded to me on the basis of the present
dissertation.
July 18, 2016

Anish Pandey
NIT Rourkela
v
Acknowledgment
My first thank is to the Almighty God, without whose blessings, I wouldn't have been
writing this “acknowledgments".
I would like to extend my heartfelt indebtedness and gratitude to Prof. Dayal R. Parhi
for his kindness in providing me an opportunity to work under his supervision and
guidance. During this period, without his endless efforts, immense knowledge, deep
patience, invaluable guidance and answers to my numerous questions, this research
would have never been possible. I am especially obliged to him for teaching me both
research and writing skills, which have been proven beneficial for my current research
and future career. He showed me different ways to approach a research problem and the
need to be persistent to accomplish any goal. It has been a great honor and pleasure for
me to do research under the supervision of Dr. Dayal R. Parhi. I am thankful to Prof.
Sunil Kumar Sarangi, Director of National Institute of Technology, for giving me an
opportunity to be a part of this institute of national importance and to work under the
supervision of Prof. Dayal R. Parhi. I am sincerely obliged to Prof. S. S. Mahapatra, Head
of the Department, Department of Mechanical Engineering, for providing me all official
and laboratory facilities during the research period. His incessant encouragement towards
research work has inspired me a lot.
I express my gratitude to Prof. P. K. Ray, Chairman DSC and DSC members for their
indebted help and valuable suggestions for the accomplishment of the dissertation. I
thank all the members of the Department of Mechanical Engineering, and the Institute,
who helped me in various ways towards the completion of my work.
I would like to thank all my friends and lab-mates for their encouragement and
understanding. Their support and lots of lovely memories with them can never be
captured in words. Finally, I thank my parents, beloved wife, my little baby Atharv, and
the entire family members for their unlimited support and strength.
July 18, 2016 Anish Pandey

NIT Rourkela Roll Number: 512ME119
vi
Abstract
The applications of the autonomous mobile robot in many fields such as industry, space,
defence and transportation, and other social sectors are growing day by day. The mobile
robot performs many tasks such as rescue operation, patrolling, disaster relief, planetary
exploration, and material handling, etc. Therefore, an intelligent mobile robot is required
that could travel autonomously in various static and dynamic environments. The present
research focuses on the design and implementation of the intelligent navigation
algorithms, which is capable of navigating a mobile robot autonomously in static as well
as dynamic environments.
Navigation and obstacle avoidance are one of the most important tasks for any mobile
robots. The primary objective of this research work is to improve the navigation accuracy
and efficiency of the mobile robot using various soft computing techniques. In this
research work, Hybrid Fuzzy (H-Fuzzy) architecture, Cascade Neuro-Fuzzy (CN-Fuzzy)
architecture, Fuzzy-Simulated Annealing (Fuzzy-SA) algorithm, Wind Driven
Optimization (WDO) algorithm, and Fuzzy-Wind Driven Optimization (Fuzzy-WDO)
algorithm have been designed and implemented to solve the navigation problems of a
mobile robot in different static and dynamic environments. The performances of these
proposed techniques are demonstrated through computer simulations using MATLAB
software and implemented in real time by using experimental mobile robots.
Furthermore, the performances of Wind Driven Optimization algorithm and Fuzzy-Wind
Driven Optimization algorithm are found to be most efficient (in terms of path length and
navigation time) as compared to rest of the techniques, which verifies the effectiveness
and efficiency of these newly built techniques for mobile robot navigation. The results
obtained from the proposed techniques are compared with other developed techniques
such as Fuzzy Logics, Genetic algorithm (GA), Neural Network, and Particle Swarm
Optimization (PSO) algorithm, etc. to prove the authenticity of the proposed developed
techniques.
Keywords: Intelligent Mobile Robot; Navigation; Hybrid Fuzzy; Cascade Neuro-Fuzzy;

Simulated Annealing algorithm; Wind Driven Optimization algorithm.
vii
Contents
Certificate of Examination ii
Supervisor’s Certificate iii
Dedication iv
Declaration of Originality v
Acknowledgements vi
Abstract vii
List of Figures xiii
List of Tables xx
Symbols & Abbreviations xxiii
1 Introduction 1-4
1.1 Background and Motivations 1
1.2 Aims and Objectives of the Proposed Research Work 2
1.3 Methodologies Applied for Proposed Research Work 2
1.4 Novelty of the Proposed Research Work 3
1.5 Outline of the dissertation 4
2 Literature Review 5-24
2.1 Introduction 5
2.2 Kinematic and Dynamic Analysis of the Wheeled Mobile Robot 7
2.3 Various Soft Computing Techniques used for Mobile Robot 8

Navigation
2.3.1 Fuzzy Logic Technique for Mobile Robot Navigation 8
2.3.2 Neural Network Technique for Mobile Robot Navigation 12
2.3.3 Neuro-Fuzzy Technique for Mobile Robot Navigation 14
viii
2.3.4 Genetic Algorithm for Mobile Robot Navigation 17
2.3.5 Simulated Annealing Algorithm for Mobile Robot 18

Navigation
2.3.6 Particle Swarm Optimization Algorithm for Mobile Robot 20

Navigation
2.3.7 Ant Colony Optimization Algorithm and Other 21

Nondeterministic Algorithms for Mobile Robot Navigation
2.3.8 Wind Driven Optimization Algorithm 23
2.4 Summary 24
3 Kinematic and Dynamic Analysis of the Nonholonomic Differential 25-33

Drive Wheeled Mobile Robot
3.1 Introduction 25
3.2 Kinematic Model of the Nonholonomic Differential Drive Two- 25

Wheeled Mobile Robot
3.3 Dynamic Model of the Nonholonomic Differential Drive Two- 32

Wheeled Mobile Robot.
3.4 Summary 33
4 Intelligent Navigation of a Mobile Robot in Static and Dynamic 34-59

Environments using Hybrid Fuzzy Architecture
4.1 Introduction 34
4.2 Hybrid Fuzzy (H-Fuzzy) Architecture 35
4.2.1 Takagi-Sugeno Type Fuzzy Logic Architecture (TFa) for 36

Goal Reaching
4.2.2 Mamdani-Type Fuzzy Logic Architecture (MFa) for Obstacle 40

Avoidance
4.3 Simulation Studies 44
4.4 Comparison with Previous Works 50
ix
4.5 Experimental Studies 53
4.5.1 Arduino Microcontroller based Wheeled Mobile Robot 53

Description
4.5.2 Experiments 54
4.6 Summary 59
5 Intelligent Navigation Control of a Mobile Robot in Unknown 60-84

Environments using Cascade Neuro-Fuzzy Architecture
5.1 Introduction 60
5.2 Cascade Neuro-Fuzzy (CN-Fuzzy) Architecture 61
5.2.1 Cascade Neural Network for Goal Reaching 62
5.2.2 Fuzzy logic architecture (FLA) for obstacle avoidance 65
5.3 Computer Simulation Results 69
5.4.1 First Comparison with Previous Works 73
5.4.2 Second Comparison with Previous Works 75
5.5 Experimental Results 77
5.5.1 Experimental Mobile Robot Description 77
5.6 Summary 84
6 Mobile Robot Navigation in Different Environments using Takagi- 85-113

Sugeno Fuzzy Controller and Simulated Annealing Algorithm
Controller
6.1 Introduction 85
6.2 Design of Sugeno-Type Fuzzy Logic Controller 87
6.3 Optimizing the Fuzzy Controller Output using Simulated Annealing 94

Algorithm (SAA)
x
6.4 Simulation Results and Discussion 97
6.6 Experimental Results and Discussion 107
6.6.1 Mobile Robot Description 107
6.7 Summary 113
7 Optimum Navigation of a Mobile Robot in the Different Environments 114-139

using Wind Driven Optimization Algorithm
7.1 Introduction 114
7.2 Path optimization using Wind Driven Optimization (WDO) algorithm 116
7.3 Computer Simulation Results and Discussion 120
7.4 Comparison with Previous Navigational Controllers 128
7.5 Experimental Results and Discussion 131
7.6 Summary 138
8 Optimum Path Planning of Mobile Robot in Unknown Static and 140-162

Dynamic Environments using Fuzzy-Wind Driven Optimization
Algorithm
8.2 Singleton Fuzzy (S-Fuzzy) Controller for the Mobile Robot 141
Navigation
8.3 Fuzzy-WDO Algorithm for the Mobile Robot Navigation 146
8.4 Simulation Results 150
8.6 Experimental Results 157
8.6.1 Khepera-III Mobile Robot Description 157
xi
8.7 Summary 162
9 Comparative Study of the Proposed Soft Computing Techniques 163-173

Applied for Mobile Robot Navigation
9.2 Simulation studies 163
9.2.1 Simulation Test1 164
9.2.2 Simulation Test2 164
9.3 Experimental studies of the developed simulations 167
9.3.1 Experimental Test1 168
9.3.2 Experimental Test2 168
9.4 Summary 173
10 Conclusion and Scope for Future Research 174-177
10.2 Important contributions 174
10.3 Conclusions 175
10.4 Scope for Future Research 177
Bibliography 178
Dissemination 196
Vitae 198
Appendix-A (Specifications of the Experimental Mobile robot) 199
xii
List of Figures
2.1 General classification of the Deterministic algorithm, Nondeterministic 6
(Stochastic) algorithm, and Evolutionary algorithm used for mobile
robot navigation
2.2 Infrared sensor based nonholonomic differential drive mobile robot 7

developed by Wang and Yang [20]
2.3 The block diagram of the fuzzy controller designed by Muthu et al. [30] 9
2.4 Behavior based fuzzy controller for mobile robot navigation and 10
obstacle avoidance developed by Algabri et al. [39]
2.5 Four-layered neural network for mobile robot navigation designed by 13

Singh and Parhi [60]
2.6 A neuro-fuzzy architecture for mobile robot navigation in 15

uncertain environments developed by Li et al. [72]
3.1 Kinematic and dynamic model of the nonholonomic differential drive 28

two-wheeled mobile robot
3.2 Robot moves straight (VL  VR ) 29
3.3 Robot turns left side (VL  VR ) 30
3.4 Robot turns right side (VL  VR ) 30
3.5 Robot rotates clockwise (VL  VR ) 31
3.6 Robot rotates anticlockwise (VL  VR ) 31
4.1 The proposed architecture of hybrid fuzzy (H-Fuzzy) logic for intelligent 36
mobile robot navigation
4.2 The basic structure of the generalized bell-shaped membership function 38
4.3 The general structure of the Takagi-Sugeno type fuzzy architecture 39

(TFa)
xiii
4.4 The membership functions of the input variables (F.O.D., L.O.D., and 39
R.O.D.)
4.5 The constant type membership function of the output variable (Turning 40
Angle)
4.6 Membership functions (i) Obstacle distances (F.O.D., L.O.D. and 43

R.O.D., respectively), (ii) Turning angle (T.A.), and (iii) Motor
velocities (Right and Left respectively)
4.7 Fuzzy logic architecture 44
4.8 Flowchart of mobile robot navigation based on H-Fuzzy architecture 46
4.9 Navigation of a mobile robot in an unknown environment using H-Fuzzy 47

architecture
4.10 Navigation of a mobile robot in an indoor environment using H-Fuzzy 47

architecture
4.11 Navigation of a mobile robot in complex environment using H-Fuzzy 48

architecture
4.12 Navigation of a mobile robot in the dynamic environment using H-Fuzzy 49

architecture
4.13 A simulation comparison results between (a) Fuzzy [163] and (b) H- 51
Fuzzy architecture
4.14 A simulation comparison results between (a) ANN [53] and (b) H-Fuzzy 52
architecture
4.15 Arduino microcontroller based experimental mobile robot 55
4.16 Sensor distribution of the experimental mobile robot 55
4.17 Experimental result of mobile robot navigation same as a simulation 56

result (shown in Figure 4.9)

result (shown in Figure 4.13 (b))
5.1 The cascade neuro-fuzzy architecture for navigation of mobile robot and 62
obstacle avoidance in unknown environments
xiv
5.2 The general structure of the cascade neural network (CNN) 64
5.3 Membership functions (i) Obstacle distances (F.O.D., L.O.D. and 68

R.O.D., respectively), (ii) Turning angle (TA), and (iii) Motor velocities
(Right and Left respectively).
5.4 Fuzzy logic architecture 69
5.5 Flowchart of the mobile robot navigation and obstacle avoidance based 70
on CN-Fuzzy architecture
5.6 Mobile robot navigation in an environment without obstacle using CN- 71

Fuzzy architecture
5.7 Mobile robot navigation in an unknown environment using CN-Fuzzy 71

architecture
5.8 Mobile robot navigation in the cluttered environment using CN-Fuzzy 72

architecture
5.9 Mobile robot navigation in the dynamic environment using CN-Fuzzy 73

architecture
5.10 Mobile robot navigation in an environment without obstacle using fuzzy 74

controller [28]
5.11 Mobile robot navigation in an environment without obstacle using CN- 75

Fuzzy architecture
5.12 Mobile robot navigation in an environment with obstacles using artificial 76

neural network [11]
5.13 Mobile robot navigation in an environment with obstacles using CN- 77

Fuzzy architecture
5.14 Experimental mobile robot 79
5.15 Sensor distribution of the experimental mobile robot 79


xv
6.1 Generalized bell-shaped membership function 88
6.2 Fuzzy membership functions for the inputs (F.O.D., R.O.D., and L.O.D.) 91
6.3 Fuzzy membership function Sugeno-Type for output variable steering 92

angle
6.4 Takagi-Sugeno type fuzzy controller 92
6.5 Rule viewer of the fuzzy controller 93
6.6 Steering angle control surface function plot 93
6.7 Fuzzy-SA controller for navigation of a mobile robot 96
6.8 Objective function value versus number of iteration number 96
6.9 The developed architecture of mobile robot navigation based on Fuzzy- 98

SA algorithm
6.10 Mobile robot navigation among the single obstacle using (a) Fuzzy 99
controller and (b) Fuzzy-SA controller
6.11 Mobile robot navigation among the many obstacles using (a) Fuzzy 100
6.12 Mobile robot navigation among the polygonal obstacles using (a) Fuzzy 101
6.13 Mobile robot navigation in the dynamic environment using Fuzzy-SA 102
controller
6.14 The graphical comparison between the (a) Martinez-Alfaro et al. [107] 105
model and (b) Proposed hybrid model
6.15 The graphical comparison between the (a) Liu et al. [37] model and (b) 106
Proposed hybrid model
6.16 Two-wheeled mobile robot 108
6.17 The arrangement of the sensors of a mobile robot 109
xvi

7.1 The architecture of mobile robot navigation based on WDO method 121
7.2 Navigation of a mobile robot using WDO technique 123
7.3 Navigation of a mobile robot using GA technique 124
7.4 Navigation of a mobile robot using PSO technique 124
7.5 Navigation of a mobile robot using WDO technique in a dynamic 125

environment
7.6 Navigation of a mobile robot using WDO technique in a cluttered 126

environment
7.7 Comparison performance graph between WDO algorithm over GA, and 126
PSO in terms of navigation path length
7.8 Comparison performance graph between WDO algorithm over GA, and 127
PSO in terms of time taken to reach the target
7.9 A simulation comparison results between GA (i) and WDO (ii) 130
7.10 A simulation comparison results between PSO (i) and WDO (ii) 131
7.11 Experimental four-wheeled real mobile robot 133
7.12 Schematic diagram of differentially steered four-wheeled mobile robot 133

environment (shown in Figure 7.2)

environment (shown in Figure 7.9 (ii))
7.15 Comparison of path length between simulation and experimental results 137
8.1 The general structure of the generalized bell-shaped membership 143

function
8.2 The structure of an S-Fuzzy controller for mobile robot navigation 144
xvii
8.3 Fuzzy membership functions for the inputs ( d f , d l , and d r ) 145
8.4 Fuzzy membership functions for the outputs ( mr , and ml ) 145
8.5 Air parcels representation of the WDO algorithm 148
8.6 Fuzzy membership functions for the inputs ( d f , d l , and d r ) after 149
optimization
8.7 Fuzzy membership functions for the outputs ( mr , and ml ) after 150
optimization
8.8 Mobile robot navigation between the obstacles using (a) S-Fuzzy and (b) 151
Fuzzy-WDO controller
8.9 Mobile robot navigation between the walls using (a) S-Fuzzy and (b) 152
8.10 Mobile robot navigation in the dynamic environment using Fuzzy-WDO 153
controller
8.11 Mobile robot navigation in an environment without obstacle using fuzzy 155
controller [45]
8.12 Mobile robot navigation in an environment without obstacle using 155

8.13 Mobile robot navigation in an environment with four obstacles using 156
fuzzy controller [45]
8.14 Mobile robot navigation in an environment with four obstacles using 156
8.15 Infrared proximity sensor distribution of Khepera-III mobile robot 158
8.16 Real time navigation of Khepera-III mobile robot between the obstacles 159
using S-Fuzzy and Fuzzy-WDO controller
8.17 Real time navigation of Khepera-III mobile robot between the walls 160
using S-Fuzzy and Fuzzy-WDO controller
9.1 Mobile robot navigation and obstacle avoidance in the simulation test1 165
using the developed soft computing techniques
xviii
9.2 Mobile robot navigation and obstacle avoidance in the simulation test2 166
using the developed soft computing techniques
9.3 Mobile robot navigation and obstacle avoidance in the experimental 170
test1 using the developed soft computing techniques
9.4 Mobile robot navigation and obstacle avoidance in the experimental 171
test2 using the developed soft computing techniques
A1 Arduino microcontroller based experimental differential drive mobile 199

robot.
xix
List of Tables
4.1 Fuzzy rule set of the Takagi-Sugeno type fuzzy logic architecture (TFa) 38
4.2 Fuzzy rule set of the Mamdani-type fuzzy logic architecture (MFa) 42
4.3 Simulation results of mobile robot navigation in the different static and 50
dynamic environments using H-Fuzzy architecture
4.4 Comparison of simulation results between Fuzzy [163] method over 53

proposed H-Fuzzy architecture
4.5 The simulation results of ANN [53] method over proposed H-Fuzzy 53
architecture in the cluttered environment
4.6 Experimental results of mobile robot navigation in the different static 58

and dynamic environments using H-Fuzzy architecture
4.7 Travelling path lengths comparison between simulation and 58

experimental results
4.8 Navigation time comparison between simulation and experimental 58

results
5.1 The different training patterns for mobile robot navigation 65
5.2 Fuzzy rule sets for navigation of mobile robot and obstacle avoidance 67
5.3 Simulation results of mobile robot navigation in the different 72

environments using CN-Fuzzy architecture
5.4 The simulation result comparison between the fuzzy controller [28] and 75
proposed CN-Fuzzy architecture
5.5 The simulation result comparison between the artificial neural network 77
[11] and proposed CN-Fuzzy architecture
5.6 Experimental results of a mobile robot navigation in the different 83


xx
results
6.1 Fuzzy control rules for mobile robot navigation using two-membership 91
functions
6.2 The result comparison between the singleton fuzzy controller and the 103
Fuzzy-SA controller
6.3 The simulation result comparison between the Martinez-Alfaro et al. 105
[107] model and Proposed hybrid model
6.4 The result comparison between the Liu et al. [37] model and proposed 107
hybrid model
6.5 Navigation path lengths between simulation and experimental results 112


results
7.1 Parameters used in WDO algorithm 122
7.2 Parameters used in GA 122
7.3 Parameters used in PSO algorithm 123
7.4 Comparison the performance of WDO algorithm over GA and PSO in 127
terms of navigation path length
7.5 Comparison the performance of WDO algorithm over GA and PSO in 128
terms of time taken to reach the target
7.6 Results of Jianguo et al. [169] method and current chapter 130
7.7 The comparison result between Deepak et al. [173] method and this 131
chapter on the optimized navigation path problem
7.8 Main specifications of the proposed prototype experimental mobile robot 134
7.9 Experimental results of mobile robot navigation in the different 137

environments using WDO algorithm
7.10 Navigation path lengths between simulation and experimental results 138
xxi
results
8.1 Fuzzy rules set 144
8.2 Adjusting parameters of the inputs before optimization 146
8.3 Adjusting parameters of the outputs before optimization 146
8.4 Adjusting parameters of the inputs after optimization 148
8.5 Adjusting parameters of the outputs after optimization 149
8.6 The simulation results of S-Fuzzy and Fuzzy-WDO controllers 154
8.7 The simulation result comparison between the fuzzy controller [45] and 157
proposed Fuzzy-WDO controller
8.8 The experimental results of S-Fuzzy and Fuzzy-WDO controllers 161


results
9.1 Simulation results of the mobile robot navigation in the test1 and test2 167
using all developed techniques
9.2 Experimental results of the mobile robot navigation in the test1 and test2 172
using all developed techniques
A1 Specifications of the experimental mobile robot 200
xxii
Symbols & Abbreviations
VR Right Wheel Linear Velocity
VL Left Wheel Linear Velocity
R Angular Velocity of Right Wheel
L Angular Velocity of Left Wheel
θ Steering Angle (Turning Angle)
C Center of Mass of a Mobile Robot
R Radius of Wheel
V Centre Linear Velocity of the Robot
 Centre Angular (Rotational) Velocity of Left Wheel
L Track Width of the Robot
m Total Mass of the Mobile Robot
I Moment of Inertia of the Robot
R Right Wheel (Motor) Torques
L Left Wheel (Motor) Torques
df Forward Obstacle Distance
dl Left Forward Obstacle Distance
dr Right Forward Obstacle Distance
mr Right Motor Velocity
ml Left Motor Velocity
TFa Takagi-Sugeno Type Fuzzy Logic Architecture
MFa Mamdani-Type Fuzzy Logic Architecture
F.O.D. Front Obstacle Distance
xxiii
L.O.D. Left Obstacle Distance
R.O.D. Right Obstacle Distance
T.A. Turning Angle
RMV Right Motor Velocity
LMV Left Motor Velocity
FLA Fuzzy logic architecture
SAA Simulated Annealing Algorithm
WDO Wind Driven Optimization
GA Genetic Algorithm
PSO Particle Swarm Optimization Algorithm
ACO Ant Colony Optimization Algorithm
ANFIS Adaptive Neuro-Fuzzy Inference System
ANN Artificial Neural Network
Note: - The symbols and abbreviations other than above have been explained in the text.
xxiv
Chapter 1
Introduction
1.1 Background and Motivations
Mobile robot is an autonomous agent capable of navigating intelligently anywhere using

sensor-actuator control techniques. The mobile robot performs various tasks such as
material handling in the industries, planetary exploration in the Mars and other planets,
and other social sectors without human intervention. Current research in the field of
mobile robotics focuses on designing and developing an intelligent algorithm or
technique, which can control the motion and orientation of the mobile robot with obstacle
avoidance/wall following competence in the static and dynamic environments. Successful
autonomous mobile robot navigation in the environment depends on its
technique/controller. Basically, during navigation, the mobile robot faces two types of
obstacles: static and dynamic. Several techniques have been applied by the various
researchers for mobile robot navigation and obstacle avoidance. According to literature
survey, it is found that the static obstacle avoidance is comparatively easy from the
dynamic obstacle avoidance. Therefore, the author is motivated to solve the static and
dynamic obstacle avoidance problem using various soft computing techniques such as
Hybrid Fuzzy (H-Fuzzy) architecture, Cascade Neuro-Fuzzy (CN-Fuzzy) architecture,
Fuzzy-Simulated Annealing (Fuzzy-SA) algorithm, Wind Driven Optimization (WDO)
algorithm, and Fuzzy-Wind Driven Optimization (Fuzzy-WDO) algorithm. The rest of
this chapter is organized as follows: Section 1.2 introduces the aims and objectives of the
proposed research work. Section 1.3 describes the methodologies applied for proposed
research work. Section 1.4 presents the novelty of the proposed research work. Finally,
Section 1.5 gives an outline of each chapter of the dissertation.
1
1.2 Aims and Objectives of the Proposed Research Work
The aims and objectives for the research towards mobile robot navigation in the static and
dynamic environments are summarized below: -
 To analyze various soft computing techniques (such as H-Fuzzy architecture, CN-

Fuzzy architecture, Fuzzy-SA algorithm, WDO algorithm, and Fuzzy-WDO) for
navigating a mobile robot from start position to goal position while avoiding static
and dynamic obstacles present in the environment.
 Integration of various sensors such as ultrasonic range finder sensors, sharp

infrared range sensors for mapping the environment cluttered with dynamic and
static obstacles.
 To design a simulated environment for carrying out the simulation exercises using
the above mentioned soft computing techniques.
 To develop experimental setup to perform the experimental exercises using the

above mentioned soft computing techniques.
1.3 Methodologies Applied for Proposed Research Work
In this research work, Hybrid Fuzzy (H-Fuzzy) architecture, Cascade Neuro-Fuzzy (CN-
Fuzzy) architecture, Fuzzy-Simulated Annealing (Fuzzy-SA) algorithm, Wind Driven
Optimization (WDO) algorithm, and Fuzzy-Wind Driven Optimization (Fuzzy-WDO)
algorithm have been designed and implemented to solve the navigation problems of a
mobile robot in different environments.
The methodologies applied for proposed research work is summarized as follows: -
 To study the various techniques applied to the mobile robot navigation in the
literature survey.
 To study the kinematic and dynamic analysis of the nonholonomic differential
drive wheeled mobile robot.
 To develop the Hybrid Fuzzy (H-Fuzzy) architecture for intelligent mobile robot
navigation and obstacle avoidance in the static and dynamic environments.
2
 To design a Cascade Neuro-Fuzzy (CN-Fuzzy) architecture to improve the
navigation and obstacle avoidance strategies of the mobile robot in various (static
and dynamic) environments.
 To integrate the Takagi-Sugeno fuzzy model with the simulated annealing
algorithm called as Fuzzy-Simulated Annealing (Fuzzy-SA) algorithm to optimize
the navigation path length of the mobile robot in the given environment.
 To apply a Wind Driven Optimization (WDO) algorithm to solve the optimal path
planning problems of a mobile robot in various simulation and experimental
environments.
 To make a hybridization of the Fuzzy-Wind Driven Optimization algorithm to
adjust and tune the input/output membership function parameters of the fuzzy
controller. This developed algorithm improves the navigation performance of the
mobile robot in the given environments and produces a smooth navigation path
within a reasonable time.
 To make a comparative study of all proposed developed techniques for checking
its strength and weakness in the various environments.
 To demonstrate the various simulation and experimental results of the proposed
techniques using the simulation and experimental setup.
1.4 Novelty of the Proposed Research Work
In literature survey, it is found that most of the researchers have applied the various soft
computing techniques for mobile robot navigation in only static environments. However,
few researchers have considered dynamic environments for mobile robot navigation. The
novelty of this dissertation is to design, analysis, and develop soft computing techniques
such as H-Fuzzy architecture, CN-Fuzzy architecture, Fuzzy-SA algorithm, WDO
algorithm, and Fuzzy-WDO algorithm for mobile robot navigation and obstacle
avoidance in the static as well as dynamic environments.
In this research work, the application of Wind Driven Optimization (WDO) algorithm
for the mobile robot navigation has been carried out. Besides, this WDO algorithm is
integrated with the fuzzy controller to adjust and optimize the antecedent and consequent
3
parameters of the fuzzy membership function and is not found during the literature
survey.
1.5 Outline of the dissertation
The rest of this dissertation is organized below: -
 Chapter-2 introduces the literature review of the kinematic and dynamic analysis
of wheeled mobile robot, and various soft computing techniques applied for
mobile robot navigation.
 Chapter-3 demonstrates the kinematic and dynamic analysis of nonholonomic
differential drive wheeled mobile robot.
 Chapter-4 presents the intelligent navigation of mobile robot in the various static
and dynamic environments using Hybrid Fuzzy (H-Fuzzy) Architecture.
 Chapter-5 describes the intelligent navigation control of mobile robot in the
various (static and dynamic) environments using Cascade Neuro-Fuzzy (CN-
Fuzzy) Architecture.
 Chapter-6 presents the mobile robot navigation among the stationary and moving
obstacle in the environments using Takagi-Sugeno Fuzzy Model and Simulated
Annealing (Fuzzy-SA) Algorithm Controller.
 Chapter-7 introduces the optimum navigation of mobile robot in the simulation
and experimental environments using Wind Driven Optimization (WDO)
Algorithm.
 Chapter-8 introduces the optimum path planning of mobile robot in unknown
static and dynamic environments using Fuzzy-Wind Driven Optimization (Fuzzy-
WDO) Algorithm.
 Chapter-9 presents the comparative study of all the proposed soft computing
techniques applied for mobile robot navigation.
 Finally, Chapter-10 describes the conclusion and scope for future research.
4
Chapter 2
Literature Review
2.1 Introduction
This chapter introduces the literature survey of the various techniques used for mobile
robot navigation. Navigation and obstacle avoidance are one of the fundamental problems
in mobile robotics, which are being solved by the various researchers in the past two
decades. The aim of navigation is to search an optimal or suboptimal path from the start
point to the goal point with obstacle avoidance competence [1]. Basically, the mobile
robot navigation has been done by the Deterministic algorithm and Nondeterministic
(Stochastic) algorithm. Nowadays, the hybridization of both the algorithms called as an
Evolutionary algorithm is being used to solve the mobile robot navigation problem.
Figure 2.1 shows the general classification of the Deterministic algorithm,
Nondeterministic (Stochastic) algorithm, and Evolutionary algorithm, which are
implemented for mobile robot navigation by various authors.
Navigation is an essential task in the field of mobile robotics, which can be classified
into two types: global navigation and local navigation. In the global navigation, the prior
knowledge of the environment should be available. Many methods have been developed
for global navigation, i.e. Voronoi graph [2, 3], Artificial potential field method [4, 5],
Dijkstra algorithm [6], Visibility graph [7], Grids [8], and Cell decomposition method [9],
and so on. In the local navigation, the robot can decide or control its motion and
orientation autonomously using equipped sensors such as ultrasonic range finder sensors,
sharp infrared range sensors, and vision (camera) sensors, etc. Fuzzy logic [10], Neural
network [11], Neuro-fuzzy [12], Genetic algorithm [13], Particle swarm optimization
algorithm [14], Ant colony optimization algorithm [15], and Simulated annealing
algorithm [16], etc. are successfully employed by various researchers to solve the local
navigation problem.
5
Rest of the chapter is organized as follows: Section 2.2 presents the literature survey
of kinematic and dynamic analysis of the wheeled mobile robots. Section 2.3 discusses
the literature review of various soft computing techniques used for mobile robot
navigation. Finally, Section 2.4 describes the summary of this literature survey.
Mobile robot navigation

algorithms
Deterministic Nondeterministic Evolutionary

algorithms algorithms algorithms
Genetic algorithm Fuzzy +

Fuzzy logic Nondeterministic
algorithm
Particle swarm Neural Network +

optimization Nondeterministic
Neural network
algorithm
Simulated
annealing
Neuro-Fuzzy
Ant colony
optimization
Wind driven
optimization
Figure 2.1: General classification of the Deterministic algorithm, Nondeterministic

(Stochastic) algorithm, and Evolutionary algorithm used for mobile robot navigation.
6
2.2 Kinematic and Dynamic Analysis of the Wheeled
Mobile Robot
The motion control problem of an autonomous wheeled mobile robot has been widely
investigated in past decades. In recent years, there has been a growing interest in the
design and development of an autonomous wheeled mobile robot using various soft
computing techniques. In [17], the authors have studied the kinematic and dynamic
constraints of a car-like mobile robot and applied it to navigation among moving
obstacles in the environments using neuro-fuzzy approaches. Abadi and Khooban [18]
have solved the trajectory tracking problem of nonholonomic wheeled mobile robots
using Random Inertia Weight Particle Swarm Optimization (RNW-PSO) based optimal
Mamdani-type fuzzy controller. The motion problem of the wheeled mobile robots on
uneven terrain has been addressed in [19]. Wang and Yang [20] have developed the
neuro-fuzzy controller for navigation of a nonholonomic differential drive mobile robot
(shown in Figure 2.2). The combination of four sharp infrared sensors is equipped on the
robot to read the obstacle distance, and this distance information is fed to the controller to
adjust the speed of two separate motors of the robot.
Figure 2.2: Infrared sensor based nonholonomic differential drive mobile robot developed
by Wang and Yang [20].
7
Wheeled mobile robots [21] have been widely used in various industrial applications,
transportation, and social sectors, etc. Martinez et al. [22] have designed the kinematics
and dynamics trajectory tracking control of the autonomous unicycle mobile robot using
type-2 fuzzy logic and genetic algorithms. An adaptive neural network based motion and
orientation control of a nonholonomic wheeled mobile robot has been presented in [23].
Liang et al. [24] have presented the kinematic modelling of the two-wheeled differential
drive mobile robot.
2.3 Various Soft Computing Techniques used for Mobile

Robot Navigation
In the past few years, many soft computing techniques are proposed by the researchers to
solve the robot navigation and obstacle avoidance problem in the various environments.
The various soft computing techniques applied for mobile robot navigation in the
different static and dynamic environments are summarized below.
2.3.1 Fuzzy Logic Technique for Mobile Robot Navigation
The concept of fuzzy logic has been introduced by Zadeh [25], which is extensively used
in many engineering applications such as mobile robotics, image processing, etc. This
method plays a vital role in the field of mobile robots. The fuzzy logic technique has been
successfully applied by many researchers to control the position and orientation of mobile
robot in the environment. Ren et al. [26] have designed an intelligent fuzzy logic
controller to solve the navigation problem of wheeled mobile robot in an unknown and
changing environment. Fuzzy logic systems are inspired by human reasoning, which
works based on perception. In [27], the authors have presented the Gradient method
based optimal Takagi-Sugeno fuzzy controller to tune the membership function
parameters, and applied it to mobile robot navigation and obstacle avoidance. Qing-yong
et al. [28] have presented the behavior-based fuzzy architecture for mobile robot
navigation in unknown environments. They have designed four basic behaviors: goal-
seeking behavior, obstacle avoidance behavior, tracking behavior, etc. for mobile robot
navigation and tested it in various simulation environments. The eight rule-based fuzzy
controllers have been designed by Boubertakh et al. [29] for obstacle avoidance and goal-
8
seeking behavior of the mobile robot. Muthu et al. [30] have presented the Atmega
microcontroller based fuzzy logic controller (Figure 2.3) for the wheeled mobile robot.
The proposed controller train the mobile robot to navigate in an environment without any
human intervention. The controller receives inputs (obstacle distance) from the group of
sensors to control the right and left motor of the mobile robot.
Figure 2.3: The block diagram of the fuzzy controller designed by Muthu et al. [30].
The sensor-based mobile robot navigation in an indoor environment using a fuzzy

logic controller has been discussed in [31-32]. Wu et al. [33] have developed the sensor
based mobile robot navigation in the narrow environment using fuzzy controller and
genetic algorithm. Where the fuzzy controller provides the initial membership function
and the genetic algorithm choose the best membership value to optimize the fuzzy
controller for mobile robot navigation. Obstacle avoidance is very important for
successful navigation of autonomous mobile robot. Samsudin et al. [34] have combined
the reinforcement learning method and genetic algorithm to optimize the fuzzy controller
for improving their performance when the mobile robot moves in an unknown
environment. Fuzzy reinforcement learning sensor-based mobile robot navigation has
been presented by Beom and Cho [35] for complex environments. Pradhan et al. [36]
have used fuzzy logic controller with different membership functions for the navigation
9
of one thousand robots in an entirely unknown environment. The authors have compared
the performance of different membership functions such as triangular, trapezoidal and
gaussian for mobile robot navigation and stated that the gaussian membership function is
more efficient for navigation. In [37], the authors have combined the fuzzy genetic
algorithm to solve the path planning and control problem of an autonomous mobile robot
(AMR) using ultrasonic range finder sensor information. Farooq et al. [38] have
presented the comparative study between the zero order Takagi-Sugeno and Mamdani-
type fuzzy logic models for mobile robot navigation and obstacle avoidance. Both the
controllers receive inputs (obstacle distance) from the left and right ultrasonic sensors to
control the left and right velocities of the motors of the mobile robot. During comparison
study, the authors have found that in terms of smoothness Mamdani-type fuzzy model
gives a better result. On the other hand, the Takagi-Sugeno fuzzy model takes less
memory space in the real-time microcontroller implementation.
Inputs Fuzzy logic controller Outputs
Target distance
Angle error
Right wheel
velocity
Front obstacle Left wheel

distance velocity
Right obstacle
distance
Left obstacle
distance
Figure 2.4: Behavior based fuzzy controller for mobile robot navigation and obstacle
avoidance developed by Algabri et al. [39].
10
Hybridization of Fuzzy and Nondeterministic Algorithm
Algabri et al. [39] have combined the fuzzy logic with other soft computing techniques
such as Genetic Algorithm (GA), Neural Networks (NN), and Particle Swarm
Optimization (PSO) to optimize the membership function parameters of the fuzzy
controller for improving the navigation performance of mobile robot. They have designed
two basic fuzzy logic behaviors: Motion to target behavior (MFLC) and obstacle
avoidance behavior (AFLC) as shown in Figure 2.4. In [40], the authors have developed
genetic-fuzzy and genetic-neural for an adaptive navigation planning of a car-like mobile
robot between dynamic obstacles. In this study, the genetic algorithm is employed to
adjust the fuzzy membership function and weight of the neural network. Fuzzy PWM
(Pulse Width Modulation) controller has been presented in the article [41] for mobile
robot navigation and obstacle avoidance in an unknown environment. Abdessemed et al.
[42] have designed an evolutionary algorithm to optimize the antecedent and consequent
parameters of the fuzzy controller, and implemented it for mobile robot path planning.
Selekwa et al. [43] have presented the fuzzy behavior controller for mobile robot
navigation in the densely obstacle populated environments. The authors have designed
two behavior control actions for navigation, namely obstacle avoidance behavior and the
goal-seeking behavior. The obstacle avoidance behavior is done by range finding sensors,
which detects the nearest obstacle distance, and the goal-seeking behavior is made by
compass measurements, which determines the direction of the goal. Pratihar et al. [44]
have developed a genetic-fuzzy technique based on a combined approach of genetic
algorithm and fuzzy logic (GA-FL) to solve the mobile robot motion planning problems
in the dynamic environments. Sensor-based wireless fuzzy controller has been designed
by Faisal et al. [45] for mobile robot navigation in the industries among the static and
dynamic objects. The two fuzzy controllers: tracking fuzzy logic control (TFLC) and
obstacle avoidance fuzzy logic control (OAFLC) are helping the robot to search collision‐
free path from the start point to goal point. Babalou and Seifiour [46] have developed the
sensor-based on-line path planning method for the mobile robot in dynamic
environments. Li et al. [47] have designed the four types of fuzzy controller: wall-
following fuzzy, corner control fuzzy, garage-parking fuzzy and parallel-parking fuzzy
for the car-like mobile robot (CLMR). The developed fuzzy controllers have been
11
implemented real-time using field-programmable gate array (FPGA) chip, and tested it in
various experimental scenarios. Li and Chang [48] have presented a real-time fuzzy target
tracking control scheme for autonomous mobile robots using infrared sensors. The
behavior-based fuzzy logic controller has been made by Dongshu et al. [49] to solve the
navigation problem of mobile robot in unknown dynamic environment. The different
fuzzy rule-based controller has been constructed to deal with different behavior and also
helps the robot to get out from the trapped situations. Antonelli et al. [50] have presented
the path-following approach for differential drive mobile robots using the fuzzy logic
technique. The designed fuzzy rules are able to emulate the human driving behavior.
Ayari et al. [51] have developed a multi-agent fuzzy logic intelligent control system,
which trains the robot to navigate autonomously in dynamic and uncertain environments.
2.3.2 Neural Network Technique for Mobile Robot Navigation
The neural network is one of the important technique for the mobile robot navigation.
This neural network technique is motivated from the human brain, which is being applied
by many researchers in the different fields such as signal and image processing, pattern
recognition, mobile robot path planning, and business, etc. Zou et al. [52] have presented
the literature survey of neural networks and its applications in mobile robotics. In [53],
the authors have combined the multi-layer feed forward artificial neural network with Q-
reinforcement learning method to construct a robust path-planning algorithm for the
mobile robot. Rai and Rai [54] have designed the Arduino Uno microcontroller-based DC
motor speed control system using the Multilayer neural network controller and
Proportional Integral Derivative (PID) controller. Patino and Carelli [55] have designed
the automatic steering controller for a mobile vehicle using neural network architecture.
Yang and Meng [56] have applied the biologically inspired neural network to generate a
collision-free path in a nonstationary environment. Biologically inspired neural network
based wall-following mobile robot has been presented by Nichols et al. [57]. Online path
planning between unknown obstacles in the environment is an interesting problem in the
field of mobile robotics. Motlagh et al. [58] have presented the target seeking, and
obstacle avoidance behaviors using neural networks and reinforcement learning. Mobile
robot navigation using hybrid neural network has been addressed by Gavrilov and Lee
[59]. Singh and Parhi [60] have designed multilayer feed forward neural network (Figure
12
2.5), which controls the steering angle of the robot autonomously in the static and
dynamic environments. The different obstacle distances are the inputs of the four-layered
neural network, and the steering angle is the output. Real-time collision-free path
planning becomes more difficult when the robot is moving in a dynamic and unstructured
environment.
Figure 2.5: Four-layered neural network for mobile robot navigation designed by Singh
and Parhi [60].
Hybridization of Neural Network and Nondeterministic Algorithm
Rossomando and Soria [61] have designed an adaptive neural network PID controller to
solve the trajectory tracking control problem of a mobile robot. Al-Jarrah et al. [62] have
described the path planning and coordination of multiple mobile robots using
probabilistic neuro-fuzzy architecture. The authors have applied leader-followers concept
to control their position and orientation in the working environment, where the follower
robots behave like a leader robot. This proposed probabilistic neuro-fuzzy architecture is
the combination of first order Sugeno fuzzy inference model and Adaptive Neuro-Fuzzy
13
Inference System (ANFIS). The fuzzy model has been used to control the linear and
angular velocities of the leader robot and the follower robots, and ANFIS is implemented
for automatic rule generation from the numerical dataset. In [63], the author has presented
a neural network-based technique for intelligent path planning and control of a mobile
robot. The two neural network controllers are applied to path planning and control. The
first neural network controller helps the robot to search free space in the environment,
and the second neural network controller trains the robot for obstacle avoidance. Glasius
et al. [64] have used Hopfield neural network for path planning and obstacle avoidance in
the complex environment. In [65], the authors have proposed type-2 fuzzy neural network
(IT2FNN) to solve the obstacle avoidance and position stabilization problems of wheeled
mobile robots. IT2FNN consists of three layers: input layer, hidden layers, and output
layer. This proposed IT2FNN has four inputs: distance between the robot and goal point,
distance between the robot and nearest obstacle, goal angle, and obstacle angle. The
outputs of the IT2FNN are linear and angular velocities of the robot. Mahmud et al. [66]
have presented the vision (camera) sensor based Kohonen-type artificial neural network
for intelligent navigation of mobile robot. Chohra et al. [67] have designed intelligent
autonomous navigation structure for a vehicle using multi-layered neural networks (NN).
Brahmi et al. [68] have solved the path planning and localization problem of mobile robot
using recurrent neural network (RNN). This RNN allows the robot to navigate
autonomously in the unknown environments. In ref. [69], the authors have controlled the
torque dynamic of nonholonomic mobile robot using neural network architecture.
2.3.3 Neuro-Fuzzy Technique for Mobile Robot Navigation
Zhu and Yang [12] have presented a neuro-fuzzy sensor based reactive navigation of
mobile robots in unknown environments. Forty-eight Fuzzy rules and two behaviors,
target seeking, and obstacle avoidance are designed using this model. A neural network
based learning techniques is developed to tune the parameters of membership functions,
which reduces the navigation path length from a start position to the end position in an
environment. Al Mutib and Mattar [70] have proposed the sensor-based navigation of
mobile robot using neuro-fuzzy architecture. The authors have used eight ultrasonic range
finder sensors for surrounding obstacle detection as the input of the neuro-fuzzy
controller for selecting the correct left and the right wheel speeds for a mobile robot.
14
Godjevac and Steele [71] have integrated the Takagi-Sugeno type fuzzy controller and
Radial basis function neural network (RBFNN) to solve the mobile robot path planning.
Where, the fuzzy logic is used to handle the uncertainty of the environment, and the
neural network is used to tune the parameters of membership functions. In [72], the
authors have constructed behaviour-based neuro-fuzzy control architecture (Figure 2.6)
for a mobile robot navigation in an unstructured environment. The neural network is used
to train the robot to reach the goal, and fuzzy architecture is integrated with it to control
the velocities of the robot.
Figure 2.6: A neuro-fuzzy architecture for mobile robot navigation in uncertain

environments developed by Li et al. [72].
Joshi and Zaveri [73] have developed a neuro-fuzzy system for reactive navigation
and control of a mobile robot in the environment with the presence of static and dynamic
obstacles. Marichal et al. [74] have designed a neuro-fuzzy sensor-actuator control
15
technique to steer the mobile robot in unknown environments. RAM based neuro-fuzzy
approach for mobile robot navigation has been presented by Zhang et al. [75]. They have
used the fuzzy rule-based controller to interpret sensory information, and neural network
controls the heading angle of the robot during navigation. Baturone et al. [76] have
designed a low-cost embedded neuro-fuzzy controller for navigation of car-like mobile
robot between the obstacles. Ma et al. [77] have used mixed soft computing techniques
like fuzzy inference system and neural network to improve the learning and decision-
making speed of a robot in unknown environments. Imen et al. [78] have applied the
Adaptive Neuro-Fuzzy Inference System (ANFIS) technique to solve the path tracking
problem of the nonholonomic wheeled mobile robots. They have used gradient descent
learning algorithm to adjust the membership function parameters of the ANFIS. In [79],
the authors have designed the two controllers: a Fuzzy Logic (FL) controller for obstacle
avoidance and Artificial Neural Network (ANN) for wall-following of the mobile robot.
Both the controllers receive inputs from the different sensors to avoid the obstacles when
the robot moves towards the desired goal. Zhao and Wang [80] have incorporated sonar
sensors with the neural network to solve the navigation problem of the autonomous
mobile robot.
Kumar and Dhama et al. [81] have integrated the neural network and fuzzy logic to
control the motion and orientation of the mobile robot in the crowded unknown
environment. In their work, the authors have used fuzzy rule-based and neural network
for goal reaching and actuator control, respectively. Song et al. [82] have designed a
heuristic fuzzy-neuro network to create a mapping between the ultrasonic sensor data and
velocity command of the robot. They have used sixteen rules to control the direction of
the mobile robot. In [83], the authors have developed a Takagi-Sugeno type recurrent
neuro fuzzy system and hybrid algorithm (genetic algorithm with particle swarm
optimization) to improve the path tracking stability of the mobile robots. The neuro-fuzzy
systems have been classified into two categories [84]: adaptive neuro-fuzzy systems
(ANFIS) and hybrid neuro-fuzzy systems. Deshpande and Bhosale [84] have discussed
the navigation of a nonholonomic wheeled mobile robot using ANFIS controller. Rusu
and Petriu et al. [85] have presented a sensor-based neuro-fuzzy controller for mobile
robot navigation in indoor environments. They have used infrared and contact sensors for
target seeking and obstacle avoidance behavior. Pothal and Parhi [86] have proposed a
16
sensor based adaptive neuro-fuzzy inference controller for navigation of single and
multiple mobile robots in the highly cluttered environment. The authors have designed
control architecture, which is able to avoid obstacle autonomously in various situations
and reach the target efficiently. Neural network integrated fuzzy controller has been
designed by Ng and Trivedi [87] for mobile robot navigation and wall-following control.
In their work, the authors have used only five rules to control the steering angle, heading
direction, and speed of the robot during wall-following. Demirli and Khoshnejad [88]
have developed sensor-based neuro-fuzzy controller for autonomous parallel parking of a
car-like mobile robot (CLMR). The proposed model received data from the sonar sensors
to control the turning angle of CLMR. Al-Mayyahi et al. [89] have applied ANFIS
technique for autonomous ground vehicle (AGV) navigation. In this work, they have
designed four ANFIS controllers to control the left and right angular velocities, and angle
between the robot and target (heading angle). In [90], the authors have designed a
navigational approach for multiple mobile robots using a neuro-fuzzy controller. The
proposed controller receives input (obstacle distance) from the array of sensors to actuate
the left and right wheel velocities of the mobile robots. Algabri et al. [91] have applied
ANFIS controller for mobile robot navigation and obstacle avoidance in an unknown
environment. The authors have presented many simulation tests using Khepera Simulator
(KiKs).
2.3.4 Genetic Algorithm for Mobile Robot Navigation
Ghorbani et al. [13] have solved the global path planning problem of a mobile robot in
the complex environment using genetic algorithm approach. Elshamli et al. [92] have
presented a genetic algorithm technique for solving the path planning problem of a
mobile robot in static and dynamic environments. Mohanta et al. [93] have designed
Petri-GA technique to optimize the navigation path length of multiple mobile robots in
the cluttered environment. Kubota et al. [94] have used the fuzzy controller to guide the
mobile robot in a static and dynamic environment, and the conventional genetic
algorithms (GAs) are integrated with it, to optimize the navigation path length. Tuncer
and Yildirim [95] have proposed a new mutation operator for a genetic algorithm (GA)
and applied it for mobile robot navigation in the dynamic environments. Moreover, the
authors have tested their developed method in various simulation environments and
17
compared it with traditional GA techniques and stated that their developed mutation
operator based GA performs better over traditional GA. In [96], the authors have
designed a genetic algorithm to choose the best membership parameters from the fuzzy
inference system and implemented it to control the steering angle of a mobile robot in the
partially unknown environment. Hu et al. [97] have designed the knowledge-based
genetic algorithm for mobile robot navigation between U-shaped obstacle and maze
environment.
Liu et al. [98] have presented the optimal path planning technique for a mobile robot
using fuzzy logic and genetic algorithm. The fuzzy controllers are applied to modify the
moving direction of the mobile robot according to the obstacle distance received from the
sensors, and genetic algorithm is used to adjust and tune membership function and rules.
Improved genetic algorithm based mobile robot navigation has been proposed by the Li et
al. [99]. The authors have done many simulation tests in the both static and dynamic
environments to show the effectiveness of the proposed algorithm. Qu et al. [100] have
developed the improved genetic algorithm instead of a conventional genetic algorithm for
global path planning of the multiple mobile robots. The advantages of the improved
genetic algorithm are capable of guiding the mobile robots efficiently from the starting
node to end node without any collision in the environment. In [101], the authors have
implemented Genetic-Fuzzy Controller (GA-FLC) to optimize and tune the Gaussian
membership function parameters for mobile robot motion control. Castillo et al. [102]
have designed Multiple Objective Genetic Algorithm (MOGA) for navigation path
optimization of the mobile robot. Arora et al. [103] have presented the single fitness
based genetic algorithm for solving the navigation problem in the dynamic environments.
They have designed a fitness function based on the Euclidean distance formula between
the robot and obstacle.
2.3.5 Simulated Annealing Algorithm for Mobile Robot Navigation
The concept of simulated annealing algorithm has come from statistical mechanics [104].
The simulated annealing is an iterative search algorithm inspired by the annealing of
metals [105]. Miao and Tian [16] have applied the heuristic method based simulated
annealing algorithm for robot path planning in the dynamic environments. The authors
have compared this proposed algorithm to the Dijkstra algorithm and stated that the
18
proposed algorithm consumes less processing time to get a solution compared to Dijkstra
algorithm. Sensor-based autonomous navigation of a mobile robot in the dynamic
environment has been presented by Chang and Song [106]. Martinez-Alfaro et al. [107]
have developed the simulated annealing and fuzzy logic for designing an automatic path
planning technique for mobile robot. The simulated annealing algorithm is used to search
a collision-free optimal trajectory between the fixed polygonal obstacles, and forty-nine
fuzzy rules are applied to adjust the velocity of the robot during navigation. Zhu et al.
[108] have presented the global path planning method for a mobile robot using Artificial
Potential Field (APF) method and Simulated Annealing Algorithm (SAA). In [109], the
authors have used SAA with fuzzy logic to adjust and optimize the antecedent and the
consequent parameters of the fuzzy membership function and applied it to solve the
optimization problem of the servo systems. Janabi-Sharifi and Vinke [110] have
addressed the local and global navigation problems in the real environment using
Artificial Potential Field method and Simulated Annealing Algorithm. Tavares et al.
[111] have discussed the off-line path planning problem of a mobile robot using SAA.
They have designed some adaptive tuning parameters to change the behavior of that
algorithm. Due to the slow convergence rate of the conventional simulated annealing
algorithm, the Liang and Xu [112] have presented a modified simulated annealing
algorithm, and applied it to mobile robot global path planning.
Nakamura and Kehtarnavaz [113] have designed an optimal fuzzy logic controller for
autonomous mobile robot navigation and hurdle avoidance using a genetic algorithm and
SAA combinatorial optimization techniques. Hussein et al. [114] have designed three
metaheuristic optimization algorithms: Tabu Search, Simulated Annealing and Genetic
Algorithm; and implemented these algorithms to improve the navigation performance of
mobile robot from the start point to goal point in an environment. Miao and Tian [115]
have presented a simulated annealing algorithm based intelligent navigational controller,
which helps the robot to search an optimal or near-optimal path in the static and dynamic
environments. Zhang et al. [116] have combined the simulated annealing algorithm and
Ant Colony Optimization (ACO) algorithm to increase the navigation speed of the mobile
robot. In [117], the authors have improved the convergence speed of the simulated
annealing algorithm using the artificial neural network and applied it to mobile robot path
planning. Synodinos and Aspragathos [118] have integrated simulated annealing
19
algorithm and artificial potential field method to rescue the robot from undesired local
minima problem during navigation. Zhao and Zu [119] have developed a Modified
Particle Swarm Optimization (MPSO) technique for mobile robot navigation in the
dynamic environment.
2.3.6 Particle Swarm Optimization Algorithm for Mobile Robot

Navigation
Particle swarm optimization (PSO) is a population-based stochastic algorithm, which is

inspired by the social behavior of bird flocks. PSO algorithm is used to find an optimal or
near optimal solution of the problem using fitness function f ( x)  f ( x1 , x2 , x3 ,.....xn ) ,
where xi is a population of the particles. Ahmadzadeh and Ghanavati [14] have presented
the PSO algorithm based navigation method for multiple mobile robots. The robots move
according to the global best (g-best) position of a particle in every iteration. To prepare an
optimal intelligent controller for an autonomous wheeled mobile robot, the Castillo et al.
[120] have designed the hybridization of an Ant Colony Optimization (ACO) algorithm
and the Particle Swarm Optimization (PSO) algorithm to optimize the membership
function of a fuzzy controller. Zhang et al. [121] have proposed the Multi-Objective
Particle Swarm Optimization Algorithm (MOPSO) to search a collision-free optimal path
in the uncertain dynamic environment. Zhang and Li [122] have presented a new
objective function for mobile robot navigation using PSO. This objective function works
based on the position of the obstacles and target in the environment. PSO algorithm has
been successfully applied by Raja and Pugazhenthi [123] to optimize the travel time of
the mobile robot in the dynamic environments. This algorithm searches the feasible path
in the environment by randomly in every iteration. Masehian and Sedighizadeh [124]
have solved the motion planning problem of the mobile robot by using multi-objective
PSO.
PSO-based optimal fuzzy controller has been designed by Wong et al. [125] to
determine the velocities of the left-wheeled motor and right-wheeled motor of the
differential drive mobile robot. Specialized particle swarm optimization algorithm has
been presented by Li et al. [126] for global optimum path planning of mobile robots. The
authors have conducted many simulation tests in the simple and complicated environment
20
to show the effectiveness of the proposed algorithm. Huang [127] has designed the
Parallel Metaheuristic Particle Swarm Optimization (PPSO) algorithm to solve the global
path planning problem of an autonomous mobile robot. The author has implemented this
PPSO algorithm in real-time using the field-programmable gate array (FPGA) chip.
Chung et al. [128] have developed PSO and fuzzy based combinatorial algorithm to
design intelligent navigation architecture for a mobile robot. They have used PSO
algorithm to escape the robot from the dead-end condition, and the fuzzy algorithm is
used to control the turn angle of a wheeled mobile robot during navigation and obstacle
avoidance. Shiltagh and Jalal [129] have investigated the application of Modified Particle
Swarm Optimization (MPSO) in the field of mobile robotics to determine a shortest
feasible path from the beginning to end in an environment between obstacles. The
developed modified PSO increases the convergence rate of the algorithms. Chatterjee and
Matsuno [130] have solved the Simultaneous Localization and Mapping (SLAM)
problem of mobile robots or vehicle using modified PSO and fuzzy evolutionary
algorithm. Juang and Chang [131] have presented an evolutionary-group-based particle-
swarm-optimization (EGPSO) for automatic learning of fuzzy system for mobile robot
navigation or wall-following control in unknown environments. In [132], the authors
have converted the robot path planning problem to the minimization problem and
designed a fitness function based on the positions of the target and obstacles in the
environment. Allawi and Abdalla [133] have proposed the sensor based PSO-fuzzy type-
2 model for the navigation of multiple mobile robots. They have used PSO algorithm to
determine the optimal input/output membership function parameters and rules for the
fuzzy type-2 controller.
2.3.7 Ant Colony Optimization Algorithm and Other Nondeterministic

Algorithms for Mobile Robot Navigation
The Ant Colony Optimization (ACO) algorithm is used by many authors for mobile robot
navigation and obstacle avoidance in the different environments. ACO is a probabilistic
algorithm proposed by Dorigo et al. [134] in 1999, which is originated from bionics.
Guan-Zheng et al. [135] have presented the modern global path planning method for a
mobile robot by applying Ant Colony System (ACS) algorithm and the Dijkstra
algorithm. Purian and Sadeghian [136] have explored the optimal path for a mobile robot
21
in an unknown dynamic environment using Ant Colony Optimization (ACO) algorithm
and fuzzy controller. This ACO algorithm searches the optimal value from the fuzzy rule
table and minimizes the distance between the start point to goal point of the mobile robot
with obstacle avoidance competence. Bi et al. [137] have designed an Ant Colony System
(ACS) to improve the path searching speed of the mobile robot in the dynamic
environment. Dong et al. [138] have presented an improved ACO algorithm for obstacle
avoidance of mobile robot in the grid environment. In [139], the authors have described
various behaviors such as goal-seeking, wall-following and obstacle avoidance for mobile
robot navigation using improved ACO algorithm. Fan et al. [140] have applied an
intensified ant colony optimization (ACO) algorithm to search an optimal path for mobile
robot between irregular obstacles in an environment. Sariff and Buniyamin [141] have
compared the performances of GA and ACO algorithm for robot path planning in the
global static environment and stated that the ACO algorithm takes less time to search an
optimal path in the environment compared to GA. Hsu et al. [142] have proposed an
improved ant colony system algorithm by including a new pheromone updating
parameter for path planning of mobile robots. Ganganath et al. [143] have designed an
off-line path planner for nonholonomic mobile robots using an ACO algorithm. Juang
and Hsu [144] have designed the reinforcement ant optimized fuzzy controller (RAOFC)
and applied it for wheeled mobile robot wall-following control under reinforcement
learning environments. The inputs of the proposed controller are range-finding sonar
sensors, and the output is a robot steering angle. The antecedent and consequent parts of
the fuzzy controller have aligned by the fuzzy type-2 clustering and ACO respectively.
Hsu and Juang [145] have designed the wall-following mobile robot using a type-2
fuzzy controller (IT2FC) and integrated it with an ACO algorithm to improve the
performance of the controller. The steering angle and moving speed of the wall-following
mobile robot has been controlled by two type-2 fuzzy controllers. In [146], the authors
have presented the navigation method of the two robots (a leader robot and a follower
robot) using fuzzy controllers (FC). They have applied continuous ant colony
optimization and particle swarm optimization (AF-CACPSO) to the control the mobile
robots to perform obstacle boundary following behavior. Hsu and Juang [147] have
adopted the multi-objective ACO for optimized the rule parameters of the fuzzy
controller (FC) for wall-following mobile robot. Chen et al. [148] have designed a scent
22
pervasion (pheromone) principle of ant (ACO) based robotic path planning in a map
environment. Hossain and Ferdousand [149] have applied Bacterial Foraging
Optimization (BFO) method for mobile robot navigation to find out shortest possible path
within the minimum time from the start position to the goal position between moving
obstacles. Liang et al. [150] have developed a bacterial foraging algorithm for making a
bio-inspired path planning strategy for a mobile robot. In the proposed model, the
behavior of bacteria is applied to search an optimal collision-free path between the start
nodes to the target node in an environment with obstacles. Brand and Yu [151] have
applied the Firefly Algorithm (Glowworm swarm optimization) to find a collision free
shortest path in the two-dimensional static and dynamic environment for a mobile robot.
They have compared this proposed algorithm to ACO algorithm and stated that the
proposed algorithm provides better results (in terms of path length and computational
cost) compared to ACO algorithm. Mohajer et al. [152] have presented a new Random
Particle Optimization Algorithm (RPOA), which is inspired by the bacterial foraging
technique, and used for local path planning for mobile robots in the dynamic and
unknown environments. The proposed algorithm randomly searches the feasible path in
the environment and avoids the moving obstacles by using the sensors.
2.3.8 Wind Driven Optimization Algorithm
Wind Driven Optimization (WDO) is a new type population-based iterative

Nondeterministic optimization algorithm, which is inspired from the earth’s atmosphere
and developed by Bayraktar et al. [153, 154]. The authors have introduced and applied
this WDO algorithm to solve the various electromagnetics optimization problems such as
synthesis of a linear antenna array, double-sided artificial magnetic conductor, and E-
shaped microstrip patch antenna. Furthermore, the authors have compared the results of
proposed algorithm to other developed algorithm such as PSO, GA, and Differential
Evolution (DE) and stated that the proposed algorithm performs better as compared to
other developed algorithm. In [155], the authors have designed an optimal high-
impedance Metasurfaces with Ultrasmall interwoven unit cells using WDO algorithm.
Kuldeep et al. [156] have applied various nature inspired optimization algorithms such as
Cuckoo Search (CS), Modified Cuckoo Search (MCS), WDO, PSO, and Artificial Bee
Colony (ABC) algorithms to optimize the 2-channel linear phase quadrature mirror filter
23
(QMF) bank design. Bhandari et al. [157] have employed CS and WDO to achieve
optimal or near-optimal threshold values of the satellite image segmentation. In an article
[158], the authors have improved the performance of WDO algorithm using Levy flights.
They have compared their performance using developed benchmark functions and stated
that in some cases Levy flights based WDO gives better results compared to standalone
WDO algorithm. Boulesnane and Meshoul [159] have presented modified WDO called as
Multi-Region Modified WDO (MR-MWDO) to improve the impact of pressure on
velocities of particles. The authors have solved the dynamic optimization problems of the
moving peaks benchmark function using developed MR-MWDO algorithm.
2.4 Summary
This chapter provides a literature survey of various techniques employed for mobile robot
navigation. After summarizing the above literature review, the major conclusions are
listed below: -
1) The various soft computing techniques e.g. Deterministic, Nondeterministic, and

Evolutionary algorithms, etc. have been applied by the researchers for mobile
robot navigation and obstacle avoidance in the different environments.
2) According to literature survey, most of the researchers have used these soft
computing techniques for mobile robot navigation and obstacle avoidance in only
static environments. However, few researchers have considered dynamic
environments for mobile robot navigation.
3) From the literature survey, it is observed that many researchers have demonstrated
only computer simulation results without implementations of physical robot.
4) Nature-inspired algorithm based mobile robot navigation and obstacle avoidance
is an important topic for the research. The hybridization of Deterministic and
Nondeterministic algorithms is also a better choice for the research.
Motivated by the aforementioned literatures, the present research work focuses on the
design and implementation of robust techniques, which can efficiently solve the
navigation and obstacle avoidance problems of the mobile robot in static and dynamic
environments.
24
Chapter 3
Kinematic and Dynamic Analysis of the

Nonholonomic Differential Drive Wheeled
Mobile Robot
3.1 Introduction
Nowadays, the wheeled mobile robots are widely used to carry out many tasks such as
planetary exploration (e.g. Mars rover), military applications (e.g. bomb disposal mobile
robot), and industrial applications (e.g. mobile manipulators), etc. Therefore, it is
important to know the kinematic and dynamic characteristics of the wheeled mobile
robots. This chapter describes the kinematic analysis of the nonholonomic differential
drive two-wheeled mobile robot. The analysis of the nonholonomic differential drive two-
wheeled mobile robot consists of two models, namely kinematic analysis and dynamic
analysis [160]. In the kinematic analysis, the motion of the robot is studied without
considering the affecting forces. On the other hand, in the dynamic analysis, we have
studied the various forces, which are responsible for this motion of the robot. The rest of
this chapter is organized as follows: Section 3.2 introduces the different kinematic
equations of the nonholonomic differential drive two-wheeled mobile robot. Section 3.3
presents the description of dynamic equations of the mobile robot. Finally, Section 3.4
depicts the summary.
3.2 Kinematic Model of the Nonholonomic Differential

Drive Two-Wheeled Mobile Robot
Figure 3.1 illustrates the kinematic and dynamic model of the nonholonomic differential
drive two-wheeled mobile robot. The term nonholonomic means the robot cannot move
25
sideways, and it moves based on the principle of rolling wheels [161]. The robot consists
of the two front driving wheels and one caster wheel for carrying the chassis. The two
separate motors have been used for driving the wheels and control the motion and
orientation of the robot. It is assumed that the robot has been made by a rigid frame, and
it is moving in a horizontal plane. In Figure 3.1, L, R, and C denotes the track width,
radius of the wheels, and center of the mass of a mobile robot, respectively. The point P is
located between the centres of the driving wheels axis. The point d is the distance
between the points P and C. The landmark (O, X, Y) shows the field navigation
environment, and (O, x, y) is the moving axis of the mobile robot. The θ is the turning
angle, which represents the orientation of the robot about an axis (O, X). The three
parameters (x, y, θ) describe the initial posture of the mobile robot, which is denoted by
q :-
q  [ x, y, ]T (3.1)
The mobile robot follows a nonholonomic constraint that means the driving wheels
purely roll without slipping. The nonholonomic constraint is written as the following
equation: -
y cos  x sin   0 (3.2)
The following equation describes the relationship between the linear velocity and angular
velocity of the wheels: -
V R (3.3)
VR  R  R (3.4)
VL  L  R (3.5)
VR  VL
 (3.6)
L
VR  VL
V (3.7)
2
The equations (3.6) and (3.7) can be written as follows: -
26
1 1 
   2 2  VR 
V
    1 
1  VL 
(3.8)
   
 L L 
The following equation calculates the moving axis (x, y) velocities of the mobile robot
and its turning (orientation) angle with respect to time ( t ): -
dx
 x  V  cos  (3.9)
dt
dy
 y  V  sin  (3.10)
dt
d
  (3.11)
dt
The kinematic equations of the two-wheeled mobile robot are as follows: -
 x  cos  0
V 
q   y    sin  0   (3.12)
 
    0 1 
The equations (3.9), (3.10) and (3.11) are updated by equations (3.6) and (3.7): -
R
x (R  L )  cos  (3.13)
2
R
y (R  L )  sin  (3.14)
2
R
 (R  L ) (3.15)
L
Kinematic model of the mobile robot is obtained by the combining of the equations
(3.13), (3.14), and (3.15): -
27
R R 
 2 cos  2
cos  
 x   
 y    R sin   
sin    R 
R
  2 2  L  (3.16)
   
 R R 

 L L 
 cos  cos  
 2 
 x   2 
 y    sin  sin   VR 
   2 2  VL 
(3.17)
   
 1 1 
 
 L L 
Figure 3.1: Kinematic and dynamic model of the nonholonomic differential drive two-
wheeled mobile robot.
28
where VR and VL are the linear velocities of the right and left wheels, respectively, which
are used as a motion command to the motors (wheels) for mobile robot navigation and
obstacle avoidance. Similarly, R and L are the angular velocities of the right and left
wheels, respectively. The V and  denote the centre linear (forward) velocity and centre
angular (rotational) velocity of the mobile robot, respectively.
The following conditions are used to control the motion and orientation of the mobile
robot: -
(1) If VL  VR , then the robot moves straight (Figure 3.2).
(2) If VL  VR , then the robot turns left side (Figure 3.3).
(3) If VL  VR , then the robot turns right side (Figure 3.4).
(4) If VL  VR , then the robot rotates (spin) clockwise (Figure 3.5).
(5) If VL  VR , then the robot rotates (spin) anticlockwise (Figure 3.6).
Figure 3.2: Robot moves straight (VL  VR ) .
29
Figure 3.3: Robot turns left side (VL  VR ) .
Figure 3.4: Robot turns right side (VL  VR ) .
30
Figure 3.5: Robot rotates clockwise (VL  VR ) .
Figure 3.6: Robot rotates anticlockwise (VL  VR ) .
31
3.3 Dynamic Model of the Nonholonomic Differential
Drive Two-Wheeled Mobile Robot
This section presents the dynamic model of the nonholonomic differential drive two-
wheeled mobile robot. Forces must be applied to the mobile robot to generate motion
[23]. These motions are used for studying the dynamic model of the mobile robot.
According to the Euler-Lagrangian formulation, the dynamic model of the nonholonomic
differential drive two-wheeled mobile robot can be described as follows [21, 23, 162]:-
M (q)q  C (q, q)q   d  B(q)  AT (q) (3.18)
where M (q) nn is a symmetric positive definite inertia matrix, C (q, q) nn is the
centripetal and Coriolis matrix,  d n1 indicates the norm-bounded unknown external
disturbance vector, B(q) nr is the input transformation matrix,  r1 is the torque
vector, AT (q) nm is the matrix associated with the constraints, and  m1 is the
vector constraint forces.
 m 0 md sin  
M (q)   0 m md cos   (3.19)
 md sin  md cos  I 
x
q   y  (3.20)
 
0 0 md cos  
 
C (q, q)  0 0 md sin   (3.21)
0 0 0 
 
 
cos  cos  
1 
B(q)   sin  sin   (3.22)
R
L L 
  
 2 2 
32
 
   R (3.23)
 L 
  sin  
A (q)   cos  
T
(3.24)
 d 
  m( x cos  y sin  ) (3.25)
where m is the total mass of the mobile robot. I is the moment of inertia of the mobile
robot.  R and  L represent the right and left wheel (motor) torques, respectively.
3.4 Summary
This chapter introduces the kinematic and dynamic equations of the nonholonomic
differential drive two-wheeled mobile robot. The major contributions of this chapter are
summarized as follows: -
 The kinematic and dynamic analysis of the mobile robot provides the knowledge
about its positions in the landmark (O, X, Y).
 Before applying soft computing techniques for mobile robot navigation and obstacle
avoidance, it is important to know the kinematic and dynamic parameters of the
wheeled mobile robot.
 Using these right and left wheel velocity equations, the motion and orientation of
the mobile robot can be controlled during navigation in the environment.
 Five conditions have been described to control the steering of the nonholonomic
differential drive mobile robot for obstacle avoidance.
33
Chapter 4
Intelligent Navigation of a Mobile Robot

in Static and Dynamic Environments
using Hybrid Fuzzy Architecture
4.1 Introduction
Navigation and obstacle avoidance are one of the most important tasks for any mobile
robot. This chapter proposes the design and implementation of a hybrid fuzzy (H-Fuzzy)
architecture for intelligent navigation of a mobile robot in the static and dynamic
environments. The proposed H-Fuzzy architecture is the combination of Takagi-Sugeno
type and Mamdani-type fuzzy logics, which helps the mobile robot to reach the goal with
obstacle avoidance. The Takagi-Sugeno type fuzzy logic architecture (TFa) is used to
assist the robot to reach the goal. The inputs of the Takagi-Sugeno type fuzzy logic
architecture are obstacle distances received from the group of sensors, and the output is
the turning angle between the robot and the goal. The Mamdani-type fuzzy logic
architecture (MFa) is integrated with the TFa to control the motor velocities of the robot.
Computer simulations are conducted through MATLAB software and implemented in
real time by using Arduino microcontroller based wheeled mobile robot. Moreover, the
successful experimental results on the actual mobile robot demonstrate the superiority of
the proposed architecture.
Over the last two-three decades, various techniques have been employed for mobile
robot navigation and obstacle avoidance. However, the most of the researchers have
applied fuzzy logic techniques for solving these types of the problem because there is a
unique feature in fuzzy technique to handle the system uncertainty without large
computation model. As compared to other soft computing techniques, the fuzzy logic
system is easy to understand. The concept of fuzzy logic has been introduced by Zadeh
34
[25], which is extensively used in many engineering applications such as mobile robotics.
The structure of fuzzy logic has been created by human knowledge, or it can be generated
from the dataset. The other soft computing techniques such as Neural Network (NN),
Genetic Algorithm (GA), Particle Swarm Optimization (PSO) algorithm, Simulated
Annealing Algorithm (SAA), and Ant Colony Optimization (ACO) algorithm have been
widely integrated with this fuzzy logic to improve its performance. For autonomous
mobile robot navigation, it is important to control the orientation and speed of the mobile
robot using sensor data. In [144, 146], the authors have been designed and implemented
an evolutionary fuzzy controller for mobile robot wall-following control.
This chapter utilizes combined fuzzy architecture, called as H-Fuzzy architecture for
intelligent mobile robot navigation. The proposed H-Fuzzy architecture aims to control
the turning angle (between the robot and goal) and motor velocities of a mobile robot
from start point to goal point with obstacle avoidance competence. This architecture is
tested in various simulation and experimental environment and is found to be a good
agreement. Rest of the chapter is organized as follows: Section 4.2 introduces the design
and implementation of the H-Fuzzy architecture for intelligent navigation of the mobile
robot in the static and dynamic environments. Section 4.3 demonstrates the computer
simulation results in different unknown environments. Section 4.4 describes the
simulation result comparison with previous works. Section 4.5 presents the experimental
results and discussion for validating the proposed controller. Finally, Section 4.6 depicts
the summary.
4.2 Hybrid Fuzzy (H-Fuzzy) Architecture
This section introduces the design and implementation of the H-Fuzzy architecture for
intelligent navigation of a mobile robot in the static and dynamic environments. Takagi-
Sugeno type fuzzy logic architecture (TFa) is used to assist the robot to reach the goal,
and the Mamdani-type fuzzy logic architecture (MFa) is used to control the right motor
velocity and left motor velocity of the mobile robot. Both the fuzzy logic architectures
(TFa and MFa) receive input (obstacle distances) from a group of sensors to control the
turning angle and motor velocities of the mobile robot during navigation. Figure 4.1
35
illustrates the proposed architecture of the hybrid fuzzy logic for intelligent mobile robot
navigation.
Figure 4.1: The proposed architecture of hybrid fuzzy (H-Fuzzy) logic for intelligent
mobile robot navigation.
4.2.1 Takagi-Sugeno Type Fuzzy Logic Architecture (TFa) for Goal

Reaching
In this section, the Takagi-Sugeno type fuzzy logic architecture (TFa) is used to assist the
mobile robot to reach the goal in the static and dynamic environments. The proposed TFa
has three inputs and single output. The TFa receives these inputs (obstacle distances)
from the front, left, and the right group of sensors of the robot. These sensors read the
obstacle from 20cm to 150cm approximately. So, the ranges of inputs are divided
between 20cm to 150cm. These inputs are denoted by F.O.D. (Front Obstacle Distance),
L.O.D. (Left Obstacle Distance), and R.O.D. (Right Obstacle Distance), respectively. The
two generalized bell-shaped (Gbell) linguistic variables, namely CLOSE and AWAY,
respectively, are selected for F.O.D., L.O.D., and R.O.D. The output of this TFa is a
36
turning angle (T.A.) between the robot and goal. The two constant type linguistic
variables NEGATIVE and POSITIVE respectively have been selected for the output, and
it is located at -90 and 90 respectively. Table 4.1 describes the fuzzy rule set of the TFa,
which helps the robot to reach the goal in the different environments. Figure 4.3
illustrates the general structure of the Takagi-Sugeno Type Fuzzy Logic Architecture
(TFa). Figures 4.4 and 4.5 show the linguistic variables of the inputs and output,
respectively. This fuzzy architecture is composed through zero-order Takagi-Sugeno
model in the following form: -
Rulen : If p1 is X i1 , p2 is X i 2 , And p3 is X i 3 THEN fi is i (4.1)
where n =1, 2, 3…8 (eight rules), the symbols p1 , p2 , and p3 are the input variables.
The X i1 , X i 2 , and X i 3 are fuzzy sets of the input variables, and  i is a real number. The
i =1, 2, because each input has two Gbell membership functions. The fuzzy set X ij uses
the following generalized bell-shaped linguistic variables: -
ij  p j  
1
2 bij
(4.2)
p j - cij
1
aij
where j =1…3 (three inputs), the cij , aij , and bij are the center, half width, and slope
controlling parameters of each generalized bell-shaped membership function,

respectively. The basic structure of the generalized bell-shaped membership function is
shown in Figure 4.2. The firing strength  i ( p) has calculated by the following function: -
n
 i ( p)   ij ( p j ) (4.3)
j 1
The defuzzification of the output variable (turning angle) is calculated by the weighted
average method: -
  ( p)  
8
 n 1 i i
fi (4.4)
  ( p)
8
n 1 i
37
Degree of Membership
Values
Figure 4.2: The basic structure of the generalized bell-shaped membership function.
Table 4.1: Fuzzy rule set of the Takagi-Sugeno type fuzzy logic architecture (TFa)
Fuzzy rules F.O.D. (cm) L.O.D. (cm) R.O.D. (cm) T. A. (degree)
1 Away Away Away Positive
2 Close Close Close Negative
3 Away Close Away Negative
4 Away Away Close Positive
5 Close Away Away Negative
6 Close Close Away Negative
7 Close Away Close Positive
8 Away Close Close Positive
38
Figure 4.3: The general structure of the Takagi-Sugeno type fuzzy architecture (TFa).
(cm)
Figure 4.4: The membership functions of the input variables (F.O.D., L.O.D., and
R.O.D.).
39
Figure 4.5: The constant type membership function of the output variable (Turning
Angle).
4.2.2 Mamdani-Type Fuzzy Logic Architecture (MFa) for Obstacle

Avoidance
This section introduces the design of Mamdani-type fuzzy logic architecture (MFa) for
mobile robot navigation and obstacle avoidance in the static and dynamic environments.
The proposed MFa is used to control the right motor velocity and left motor velocity of
the mobile robot. The MFa has four inputs and two outputs. The MFa receives first three
inputs (obstacle distance) from the group of sensors of the mobile robot. The first three
inputs are denoted by F.O.D., L.O.D., and R.O.D., respectively. The fourth input is the
turning angle (goal angle) between the robot and goal, which is received from the TFa.
The outputs of the MFa are the velocities of the motors of a robot. The outputs are
addressed by RMV (Right Motor Velocity) and LMV (Left Motor Velocity),
respectively. The range of first three inputs is divided into two linguistic variables,
namely CLOSE and AWAY, respectively, and it is located at 20cm to 150cm. The two
linguistic variables NEGATIVE and POSITIVE, respectively, are used for turning angle
(T.A.). The range of outputs is divided into two linguistic variables, namely LOW and
HIGH, respectively. These outputs are located at 6.7cm/sec to 16.7cm/sec. The two
40
generalized bell-shaped (Gbell) membership functions are selected for inputs and outputs.
Figure 4.6 shows the input and output membership functions of the MFa. Figure 4.7
illustrates the fuzzy logic architecture. Table 4.2 describes the fuzzy rule set of the MFa.
The MFa is composed through Mamdani-type fuzzy model in the following form: -
Rulem : IF x1 is Aj1 , x2 is Aj 2 , x3 is Aj 3 , & x4 is Aj 4 THEN y1 is B j1 & y2 is B j 2 (4.5)
where m =1, 2, 3…12 (twelve rules), the x1 , x2 , x3 , and x4 are the input variables.
Similarly, y1 and y2 are the output variables. The Aj1 , Aj 2 , Aj 3 , and Aj 4 are the fuzzy
sets of the input variables. Similarly, B j1 and B j 2 are the fuzzy sets of the output
variables. The j =1, 2, because each input and output have two Gbell membership
functions. The fuzzy set (inputs and outputs) uses the following Gbell membership
function: -
1
 jk ( xk ; a, b, c)  2 b jk
(4.6)
xk - c jk
1
a jk
1
 jl ( yl ; a, b, c)  2 b jl
(4.7)
yl - c jl
1
a jl
where k =1…4 (four inputs), and l =1, 2 (two outputs). The symbols a jk , b jk , and c jk are
adjusting parameters of the Gbell membership function (see the Figure 4.2); called as the
half width, slope control, and center, respectively.
The defuzzification of the output variables ( y1 and y2 ) are accomplished by the weighted
average method: -
 ( x1 )   j 2 ( x2 )   j 3 ( x3 )   j 4 ( x4 )   y1
12
j1
y1  m 1
(4.8)
 ( x1 )   j 2 ( x2 )   j 3 ( x3 )   j 4 ( x4 ) 
12
j1
m 1
41
 ( x1 )   j 2 ( x2 )   j 3 ( x3 )   j 4 ( x4 )   y2
12
j1
y2  m 1
(4.9)
 ( x1 )   j 2 ( x2 )   j 3 ( x3 )   j 4 ( x4 ) 
12
j1
m 1
Table 4.2: Fuzzy rule set of the Mamdani-type fuzzy logic architecture (MFa)
Fuzzy F.O.D. L.O.D. R.O.D. T. A. RMV LMV

rules (cm) (cm) (cm) (degree) (cm/sec) (cm/sec)
1 Away Away Away Positive High Low
2 Away Away Away Negative Low High
3 Close Close Close Negative Low High
4 Close Close Close Positive High Low
5 Away Close Away Negative Low High
6 Away Away Close Positive High Low
7 Close Away Away Negative Low High
8 Close Away Away Positive High Low
9 Close Close Away Negative Low High
10 Close Away Close Positive High Low
11 Away Close Close Positive High Low
12 Away Close Close Negative Low High
42
(i) (ii)
(cm) (degree)
(iii)
(cm/sec)
Figure 4.6: Membership functions (i) Obstacle distances (F.O.D., L.O.D. and R.O.D.,
respectively), (ii) Turning angle (T.A.), and (iii) Motor velocities (Right and Left,
respectively)
43
Figure 4.7: Fuzzy logic architecture.
4.3 Simulation Studies
This section illustrates the successful simulation results of the mobile robot navigation in
a different type of environments. The simulations are conducted through the MATLAB
software on the HP desktop and implemented in real time by using Arduino
microcontroller based experimental mobile robot. The developed flowchart of mobile
robot navigation based on H-Fuzzy architecture is given in Figure 4.8. Figures 4.9 and
4.10 show the simulation results of the mobile robot navigation in the unknown and
indoor environment using H-Fuzzy architecture, respectively. In the simulation results, it
is assumed that the position of the starting point and goal point is known, but the
positions of all the obstacles in the environment are unknown for the robot. The width
and height of the environment are 250cm and 250cm, respectively. When the obstacles
are detected on the front, left, and right side of the robot, then the H-Fuzzy architecture is
activated, and the robot adjusts its motor velocities (left and right) according to the
architecture outputs.
44
Moreover, Figure 4.11 shows the mobile robot navigation in a typical U-shape
complex environment. In the simulation test, the robot starts from the initial coordinate
(10cm, 10cm) to the goal coordinate (230cm, 220cm) and there is a U-shaped complex
obstacle located in the middle of the environment. It can be seen that from the simulation,
when the robot comes near to the U-shaped obstacle, it turns right sharply and follows the
wall to protect from the collision, and reach the goal successfully by avoiding the
obstacle.
Figure 4.12 illustrates the simulation result of mobile robot navigation between the
two moving obstacles. In the simulation, the first cyan color circular shape obstacle is
moving from the coordinates (200cm, 150cm) to (59cm, 125cm) with the moving angle
190° from right to left. The second red color rectangular shape obstacle is moving from
coordinates (150cm, 210cm) to (150cm, 103cm) with the moving angle 270° from up to
down. From the simulation result, when the moving obstacles are near to the mobile
robot, then the H-Fuzzy architecture is activated, and the robot turns right, i.e. the rate of
Right Motor Velocity (RMV) decreases less than the rate of Left Motor Velocity (LMV),
respectively. Table 4.3 illustrates the simulation results of mobile robot navigation in the
different static and dynamic environments using H-Fuzzy architecture.
45
Begin
Initialize the Position of Mobile Robot and Goal
Mobile Robot Moves Towards the Goal
YES
Reach Goal Quit
NO
NO Using Sensors Reading If

Robot Near to Obstacles
YES
H-Fuzzy Architecture Active and Gives Appropriate Motor Velocities

Command
Robot Moves According to Velocities Provided by H-Fuzzy Architecture
YES NO
Obstacles Avoided
Figure 4.8: Flowchart of mobile robot navigation based on H-Fuzzy architecture.
46
(cm)
(cm)
Figure 4.9: Navigation of a mobile robot in an unknown environment using H-Fuzzy

architecture.
(cm)
(cm)
Figure 4.10: Navigation of a mobile robot in an indoor environment using H-Fuzzy

architecture.
47
(cm)
(cm)
Figure 4.11: Navigation of a mobile robot in complex environment using H-Fuzzy

architecture.
48
(cm)
(cm)
(cm) (cm)
(cm)
(cm)
(cm) (cm)
Figure 4.12: Navigation of a mobile robot in the dynamic environment using H-Fuzzy
architecture.
49
Table 4.3: Simulation results of mobile robot navigation in the different static and
Travelling path Navigation time

Figure no. Environment type
length (cm) (sec)
Figure 4.9 Unknown environment 73 9.2
Figure 4.10 Indoor environment 47 6.1
Figure 4.11 Complex environment 125 16.6
Figure 4.12 Dynamic environment 102 13.1
4.4 Comparison with Previous Works
This section describes the comparative study of proposed H-Fuzzy architecture over
Fuzzy [163] and ANN (Artificial Neural Network) [53] methods in the simulation mode.
In an article [163], the Cherroun and Boumehraz have prepared a Takagi-Sugeno type
fuzzy behaviour based control system, which helps to navigate a tricycle type mobile
robot autonomously in the unknown crowded environment. In [53], the authors have
combined the multi-layer feed forward artificial neural network with Q-reinforcement
learning method to construct a robust path-planning algorithm for the mobile robot.
The performance of this proposed architecture is evaluated by the path length and
travelling path smoothness. Figure 4.13 and Figure 4.14 demonstrates the graphical
comparison between the path made by Fuzzy [163] and ANN [53] over current developed
H-Fuzzy architecture for the same path planning problems.
From the simulation analysis, it has been seen that the proposed architecture provides
better trajectories in terms of path length and smoothness as compared to previous
methods, and also it can efficiently drive the mobile robot in an optimal path in different
environments. Tables 4.4 and 4.5 illustrate the path covered (in cm) by the robot to reach
the goal using proposed architecture and previous methods [163, 53]. The centimeter
measurements are taken on the proportional basis.
50
(a) Fuzzy
(b) H-Fuzzy architecture

(m)
(m)
Figure 4.13: A simulation comparison results between (a) Fuzzy [163] and (b) H-Fuzzy
architecture.
51
(a) ANN
Y-AXIS (cm)
X-AXIS (cm)
(b) H-Fuzzy architecture

Y-AXIS (cm)
X-AXIS (cm)
Figure 4.14: A simulation comparison results between (a) ANN [53] and (b) H-Fuzzy
architecture.
52
Table 4.4: Comparison of simulation results between Fuzzy [163] method over proposed
H-Fuzzy architecture
Navigation path length

Figure no. Method
(cm)
Figure 4.13 (a) Fuzzy [163] 91
Figure 4.13 (b) H-Fuzzy architecture 84
Table 4.5: The simulation results of ANN [53] method over proposed H-Fuzzy
architecture in the cluttered environment

Figure no. Method
(cm)
Figure 4.14 (a) ANN [53] 74
Figure 4.14 (b) H-Fuzzy architecture 69
4.5 Experimental Studies
4.5.1 Arduino Microcontroller based Wheeled Mobile Robot

Description
This section presents the description of the Arduino microcontroller based two-wheeled
differential drive experimental mobile robot (see the Figure 4.15). The mobile robot has
two wheels, which is connected to two separate DC geared motors. The motion and
orientation of the robot are controlled by two independent DC geared motors, which
provides the necessary torque to all the driving wheels. The width of the robot plate is
23cm, and the track width and height of robot are 30cm and 8cm, respectively. The
mobile robot is equipped with one sharp infrared range sensor on the front side, and the
two ultrasonic range finder sensors fitted on the left and right side of the robot, as shown
in Figure 4.16. The sharp infrared range sensor reads obstacles up to 150cm, and the
ultrasonic sensor reads obstacles from 2cm to 4m approximately. In this study, the
53
minimum and maximum velocities of the wheeled mobile robot are between the 6.7-
16.7cm/sec.
4.5.2 Experiments
The experiments are conducted using the Arduino microcontroller based mobile robot in
unknown environments. Figures 4.17 and 4.18 show the real time motion and orientation
of the experimental mobile robot in the two different environments. The width and height
of the platform are 250cm and 250cm, respectively. In the experimental results, it is
assumed that the position of the start point and goal point are known, but the positions of
all the obstacles in the environment are unknown for the robot. The proposed architecture
receives input (obstacle distances) from the front, left, and the right group of the sensors
(Figure 4.16) to control the turning angle and motor velocities of the mobile robot during
navigation. If the left obstacle is near to the mobile robot, then the robot turns right, i.e.
the velocity of the right motor is less than the velocity of left motor. Similarly, if the right
obstacle is near to the mobile robot, then the robot turns left, i.e. the velocity of the right
motor is more than the velocity of left motor.
Table 4.6 shows the experimental path length and time taken by the mobile robot to
reach the goal using H-Fuzzy architecture in two different environments. Tables 4.7 and
4.8 illustrate the travelling path length and the navigation time comparison between the
simulation and experiments. In the comparison study between the simulation and
experiments, it is observed that some errors have been found, these are happened due to
slippage and friction during real time experiment.
54
Figure 4.15: Arduino microcontroller based experimental mobile robot.
Figure 4.16: Sensor distribution of the experimental mobile robot.
55
Y-AXIS
X-AXIS
Figure 4.17: Experimental result of mobile robot navigation same as a simulation result
(shown in Figure 4.9).
56
Y-AXIS
X-AXIS
(shown in Figure 4.13 (b)).
57
Table 4.6: Experimental results of mobile robot navigation in the different static and

length (cm) (sec)
Table 4.7: Travelling path lengths comparison between simulation and experimental
results
Error between
Travelling path length (cm)
Figure no. (Simulation simulation and
and experimental res.) Simulation Experimental experimental
result result result
Figures 4.9 and 4.17 73 78 6.41%
Figures 4.13 (b) and 4.18 84 90 6.67%
Table 4.8: Navigation time comparison between simulation and experimental results
Error between
Navigation time (sec)
Figures 4.9 and 4.17 9.2 9.9 7.07%
Figures 4.13 (b) and 4.18 10.8 11.6 6.89%
58
4.6 Summary
This chapter introduces the hybrid fuzzy (H-Fuzzy) architecture for intelligent navigation
of a mobile robot. The proposed H-Fuzzy architecture has been designed for two
behaviors, i.e. goal reaching and obstacle avoidance.
Takagi-Sugeno type fuzzy logic architecture (TFa) has been used to assist the robot to
reach the goal. Mamdani-type fuzzy logic architecture (MFa) has been applied to control
the right motor velocity and left motor velocity of the mobile robot.
Furthermore, the proposed architecture generates better results (in terms of path
length) as compared to previous models [163] and [53], which verifies the superiority of
Moreover, the simulation and experimental studies demonstrate that the proposed
architecture efficiently drives the mobile robot in the different static and dynamic
environments such as unknown, indoor and complex environments. In the comparison
study between the simulation and experiment results errors are recorded, and the errors
are found due to the effect of slippage and friction between the wheels of the robot and
surface during navigation in real time mode.
During experiment utmost care has been taken to minimize the slippage and friction
between the wheels and surface. Still the effect of slippage and friction are unavoidable,
and errors are recorded during the comparison of the results for travelling path length
(6.54%) and for navigation time (6.98%).
59
Chapter 5
Intelligent Navigation Control of a Mobile

Robot in Unknown Environments using
Cascade Neuro-Fuzzy Architecture
5.1 Introduction
Real-time navigation in the partially unknown environment is an interesting task for

mobile robotics. This chapter presents the cascade neuro-fuzzy (CN-Fuzzy) architecture
for intelligent navigation control of a mobile robot in an unknown environment filled
with obstacles. The array of ultrasonic range finder sensors and sharp infrared range
sensors are used to read the front, left and right obstacle distances. The cascade neural
network is used to train the robot to reach the goal. Its inputs are the different obstacle
distance received from the sensors. The output of the neural network is a turning angle
between the robot and goal. The fuzzy architecture is integrated with the cascade neural
network to control the velocities of the robot. Successful simulation and experimental
results verify the effectiveness of the proposed architecture in both static and dynamic
environments.
The applications of the intelligent robot in many fields such as industry, space,
agriculture, defence and transportation, and other social sectors are growing day by day.
The mobile robot performs many tasks such as rescue operation, patrolling, underwater
exploration, disaster relief and planetary exploration, etc. Therefore, the author is trying
to put the effort in the field of the intelligent robot using CN-Fuzzy architecture, which
can avoid the obstacle autonomously and reach the goal safely in the given environment.
Autonomous mobile robot navigation is one of the challenging tasks for any soft
computing techniques. Fuzzy logic and neural network have been widely used for mobile
robot navigation and control because these methods are capable of handling the system
60
uncertainty. Generally, the fuzzy logic is the combination of fuzzy rules and membership
functions (inputs and outputs), which are constructed by human knowledge. The neural
network with fuzzy logic [77] improves the decision speed of the mobile robot for target
seeking and obstacle avoidance.
Cascade neural network (CNN) is similar to feed forward neural network (FNN).
Both neural networks use back propagation algorithm for updating the weights and biases
[164]. This chapter describes the cascade neural network based fuzzy architecture for
mobile-robot navigation in unknown environments. The cascade neural network is used
to train the robot to reach the goal. Its inputs are different obstacle distance received from
the sensors. The output of the neural network is a turning angle between the robot and
goal. The fuzzy logic architecture is used to control the right motor velocity and left
motor velocity of the mobile robot. In the last two decades, many researchers have
implemented different neuro-fuzzy techniques for solving the navigation problem of the
mobile robot. Motivated by the above literature survey, the primary objective of this
chapter is to improve the path planning accuracy and efficiency of the mobile robot using
the cascade neuro-fuzzy controller. The remainder of this chapter is structured as follows:
Section 5.2 introduces the design and implementation of the CN-Fuzzy architecture for
navigation of mobile robot and obstacle avoidance in unknown environments. Section 5.3
demonstrates the computer simulation results in different unknown environments. Section
5.4 describes the simulation result comparison with previous works. Section 5.5 presents
the experimental results and discussion for validating the proposed controller. Finally,
Section 5.6 depicts the summary.
5.2 Cascade Neuro-Fuzzy (CN-Fuzzy) Architecture
This section introduces the design and implementation of the CN-Fuzzy architecture for
navigation of mobile robot and obstacle avoidance in unknown environments. The
cascade neural network is used to train the robot to reach the goal in the environment, and
the fuzzy logic architecture is used to control the right motor velocity and left motor
velocity of the mobile robot. Figure 5.1 shows the proposed architecture of CN-Fuzzy for
navigation of mobile robot and obstacle avoidance in unknown environments.
61
Figure 5.1: The cascade neuro-fuzzy architecture for navigation of mobile robot and
obstacle avoidance in unknown environments.
5.2.1 Cascade Neural Network for Goal Reaching
The neural network is one of the important technique for the mobile robot navigation. In
this section, the cascade neural network (CNN) is used to train the robot to reach the goal
in the environment. The neural network is the combination of many layers such as input
layer, hidden (intermediate) layers, and the output layer; all the layers are connected with
each other by the neurons. The CNN is the similar to the feed forward neural network
(FNN). Both CNN and FNN use back propagation algorithm for updating the weights and
biases. The two back propagation algorithms, namely Levenberg-Marquardt (LM) and
Bayesian regularization (BR) are used to adjust the network weights and biases. Figure
62
5.2 illustrates the general structure of a cascade neural network (CNN). In Figure 5.2, u ,
w , b , and v addresses the input variables, synaptic weights, neuron bias, and output
variable, respectively.
The inputs of the CNN are the obstacle distance received from the various sensors,
and the output of the CNN is a turning angle between the robot and goal. Table 5.1
describes the different training patterns for the cascade neural network, which helps the
robot to reach the goal in the environment. The proposed CNN uses three inputs, two
hidden layers (six and four neurons, respectively) and single output layer for the mobile
robot navigation. The CNN has three inputs: F.O.D. (Front Obstacle Distance), L.O.D.
(Left Obstacle Distance), and R.O.D. (Right Obstacle Distance), respectively. The output
of this CNN is a turning angle (T.A.) between the robot and goal. The input and output of
the CNN can be written as follows: -
Input layer (first layer):
qi[1]  ui (5.1)
where i = 1, 2, 3. (Three inputs F.O.D., L.O.D., and R.O.D., respectively)

Two hidden layers (second and third):
qt[ s ]    NETt[ s ]  (5.2)
NETt[ s ]   ( wt[si ]  qi[ s 1]  bt[ s ] ) (5.3)

i
where s = 2, 3. (Second and third layers)

Output layer (fourth layer):
v( p )  q[4]    NET [4]  (5.4)
NET [4]   (wi[4]  qi[3]  bi[4] ) (5.5)

i
where ui is the input variables, v( p ) is the predicted output variable (turning angle). The
wt[si ] is the synaptic weight on connection joining the i th neuron in the layer [ s  1] to the
63
t th neuron in the layer [ s ] ; bt[ s ] is a bias of the t th neuron in the layer [ s ] , and t  d  is
the Log-sigmoid transfer function.
1
 (d )  (5.6)
1  exp(d ) 
The proposed CNN is verified through the mean squared error (MSE) and root mean
square error (RMSE) method: -
 r  v( a )  v( p ) 2 
MSE (%)   1    100 (5.7)
  r  
1  r  v( a )  v( p )  
2
RMSE (%)  
r  1  v( a ) 
  100

(5.8)
 
where v( a ) is the actual output variable, v( p ) is the predicted (network) output variable,
and r is the number of observations.
Figure 5.2: The general structure of the cascade neural network (CNN).
64
Table 5.1: The different training patterns for mobile robot navigation
F.O.D. (cm) L.O.D. (cm) R.O.D. (cm) T.A. (degree) Turning direction
20 115 20 74.3 Left
20 20 150 -65.9 Right
125 25 150 -70.4 Right
25 75 50 55 Left
40 120 60 59.4 Left
25 150 100 72.8 Left
25 50 120 -22.9 Right
22 25 22 73.4 Left
50 25 25 0 Straight
20 27 27 77 Left
100 28 25 0 Straight
25 21 22 77.2 Left
150 25 115 -70.5 Right
150 20 25 0 Straight
150 100 100 -70.4 Right
5.2.2 Fuzzy logic architecture (FLA) for obstacle avoidance
This section describes the design of Mamdani-type fuzzy logic architecture for navigation
of mobile robot and obstacle avoidance in unknown environments. The fuzzy logic
architecture (FLA) is used to control the right motor velocity and left motor velocity of
the mobile robot. The proposed FLA has four inputs and two outputs. The FLA receives
first three inputs (obstacle distance) from the various sensors of the mobile robot. The
first three inputs are denoted by F.O.D., L.O.D., and R.O.D., respectively. The fourth
input is the turning angle (goal angle) between the robot and goal, and which is received
65
from the CNN. The outputs of the FLA are the velocities of the motors of robot. The
outputs are addressed by RMV (Right Motor velocity) and LMV (Left Motor velocity),
respectively. The range of first three inputs is divided into two linguistic variables,
namely CLOSE and AWAY, respectively, and it is located between 20cm to 150cm. The
two linguistic variables NEGATIVE and POSITIVE, respectively, are used for turning
angle. The range of outputs is divided into two linguistic variables, namely LOW and
HIGH, respectively. The two generalized bell-shaped (Gbell) membership functions are
used for inputs and outputs. Figure 5.3 shows the input and output variables of the FLA.
Figure 5.4 illustrates the fuzzy logic architecture. The fuzzy rule set of the FLA is
described in Table 5.2. The FLA is composed through Mamdani-type fuzzy model in the
following form: -
Rulem : IF x1 is Aj1 , x2 is Aj 2 , x3 is Aj 3 , & x4 is Aj 4 THEN y1 is B j1 & y2 is B j 2 (5.9)
where m =1, 2, 3…12 (twelve rules), the x1 , x2 , x3 , and x4 are the input variables.
Similarly, y1 and y2 are the output variables. The Aj1 , Aj 2 , Aj 3 , and Aj 4 are the fuzzy
sets of the input variables. Similarly, B j1 and B j 2 are the fuzzy sets of the output
variables. The j =1, 2, because each input and output have two Gbell membership
functions. The fuzzy set (inputs and outputs) uses the following Gbell membership
function: -
1
 jk ( xk ; a, b, c)  2 b jk
(5.10)
xk - c jk
1
a jk
1
 jl ( yl ; a, b, c)  2 b jl
(5.11)
yl - c jl
1
a jl
where k =1…4 (four inputs), and l =1, 2 (two outputs). The symbols a , b , and c are
adjusting parameters of the Gbell membership function; called as the half width, slope
control, and center respectively.
The defuzzification of the output variables ( y1 and y2 ) are accomplished by the weighted
66
average method: -
 ( x1 )   j 2 ( x2 )   j 3 ( x3 )   j 4 ( x4 )   y1
12
j1
y1  m 1
(5.12)
 ( x1 )   j 2 ( x2 )   j 3 ( x3 )   j 4 ( x4 ) 
12
j1
m 1
 ( x1 )   j 2 ( x2 )   j 3 ( x3 )   j 4 ( x4 )   y2
12
j1
y2  m 1
(5.13)
 ( x1 )   j 2 ( x2 )   j 3 ( x3 )   j 4 ( x4 ) 
12
j1
m 1
Table 5.2: Fuzzy rule sets for navigation of mobile robot and obstacle avoidance
Fuzzy F.O.D. L.O.D. R.O.D. T. A. RMV LMV

rules (cm) (cm) (cm) (degree) (cm/sec) (cm/sec)
1 Away Away Away Positive High Low
2 Away Away Away Negative Low High
3 Close Close Close Negative Low High
4 Close Close Close Positive High Low
5 Away Close Away Negative Low High
6 Away Away Close Positive High Low
7 Close Away Away Negative Low High
8 Close Away Away Positive High Low
9 Close Close Away Negative Low High
10 Close Away Close Positive High Low
11 Away Close Close Positive High Low
12 Away Close Close Negative Low High
67
(i) (ii)
(cm) Degree of Membership (degree)
(iii)
(cm/sec)
Figure 5.3: Membership functions (i) Obstacle distances (F.O.D., L.O.D. and R.O.D.,
respectively), (ii) Turning angle (TA), and (iii) Motor velocities (Right and Left
respectively).
68
Figure 5.4: Fuzzy logic architecture.
5.3 Computer Simulation Results
This section presents the computer simulation results using CN-Fuzzy architecture in the
different unknown static and dynamic environments. The simulations have done using
MATLAB software on the HP 3.40 GHz processor. Figure 5.5 illustrates the developed
flowchart of mobile robot navigation and obstacle avoidance based on CN-Fuzzy
architecture. Figures 5.6 to 5.9 show the mobile robot navigation trajectories in the
different static and dynamic environments. In the simulation results, it is assumed that the
position of the start point and goal point are known. But the positions of all the obstacles
in the environment are unknown for the robot. The dimension of the environments is
300cm width and 300cm height. A minimum threshold distance is fixed between the
robot and the obstacle. Now if the robot detects the obstacle in the threshold range, then
the proposed architecture estimates the desired turning direction of a mobile robot. Table
5.3 illustrates the navigation path length and time taken by the robot in the various
unknown environments.
69
Start
Initialize the Position of Mobile Robot and

Goal
Mobile Robot Navigate Towards the Goal
YES
Reach Goal Stop
NO
NO Meet Obstacles During

Navigation
YES
CN-Fuzzy Architecture Active and Provides Right and

Left Motor Velocities Command
Robot Moves According to Velocities Given by CN-Fuzzy

Architecture
YES NO
Obstacles Avoided
Figure 5.5: Flowchart of the mobile robot navigation and obstacle avoidance based on
CN-Fuzzy architecture.
70
(cm)
(cm)
Figure 5.6: Mobile robot navigation in an environment without obstacle using CN-Fuzzy
architecture.
(cm)
(cm)
Figure 5.7: Mobile robot navigation in an unknown environment using CN-Fuzzy

architecture.
71
(cm)
(cm)
Figure 5.8: Mobile robot navigation in the cluttered environment using CN-Fuzzy
architecture.
Table 5.3: Simulation results of mobile robot navigation in the different environments
using CN-Fuzzy architecture

length (cm) (sec)
Figure 5.6 Without obstacle 103 11.6
Figure 5.8 Cluttered environment 120 13.4
Figure 5.9 Dynamic environment 77 8.6
72
(cm)
(cm)
(cm) (cm)
(cm)
(cm)
(cm) (cm)
Figure 5.9: Mobile robot navigation in the dynamic environment using CN-Fuzzy
architecture.
This section describes the computer simulation result comparison between the previous
models [28, 11] and proposed CN-Fuzzy architecture in the same environment.
5.4.1 First Comparison with Previous Works

In article [28], the authors have designed goal-seeking, obstacle avoidance behavior, and
73
other behavior for mobile robot navigation by using fuzzy controller. Figures 5.10 and
5.11 illustrate the mobile robot navigation in the same environment without obstacle
using fuzzy controller [28] and CN-Fuzzy architecture, respectively. From simulation
result, it can be clearly seen that the robot covers shorter distance to reach the goal using
proposed architecture as compared to previous model [28]. Table 5.4 shows the path
covered by the robot to reach the goal using fuzzy controller [28] and proposed CN-
Fuzzy architecture. The centimetre measurements are taken on the proportional basis.
Y-AXIS (m)
X-AXIS (m)
Figure 5.10: Mobile robot navigation in an environment without obstacle using fuzzy
controller [28].
74
(m)
(m)
Figure 5.11: Mobile robot navigation in an environment without obstacle using CN-
Fuzzy architecture.
Table 5.4: The simulation result comparison between the fuzzy controller [28] and
proposed CN-Fuzzy architecture

Figure no. Method
(cm)
Figure 5.10 Fuzzy [28] 51
Figure 5.11 CN-Fuzzy architecture 46
5.4.2 Second Comparison with Previous Works

In this section, the simulation result comparison has been made between the previous
controller [11] and proposed CN-Fuzzy architecture in the same environment with the
obstacles. In [11], the authors have discussed the motion and path planning of a car-like
wheeled mobile robot between the stationary obstacles using backpropagation artificial
75
neural network. Figure 5.12 shows the mobile robot navigation in an environment with
obstacles using artificial neural network [11]. Figure 5.13 presents the path covered by
the robot using proposed CN-Fuzzy architecture in the same environment. From the
Figures 5.12 and 5.13, it is observed that the proposed architecture avoid the obstacles
with some shorter distance or minimum steering as compared to previous model [11].
Table 5.5 illustrates the path traced (in cm) by the robot to reach the goal using proposed
architecture and previous model [11]. The centimetre measurements are taken on the
proportional basis.
Y-AXIS (cm)
X-AXIS (cm)
Figure 5.12: Mobile robot navigation in an environment with obstacles using artificial
neural network [11].
76
Y-AXIS (cm)
X-AXIS (cm)
Figure 5.13: Mobile robot navigation in an environment with obstacles using CN-Fuzzy
architecture.
Table 5.5: The simulation result comparison between the artificial neural network [11]
and proposed CN-Fuzzy architecture

Figure no. Method
(cm)
Figure 5.12 Artificial neural network [11] 87
Figure 5.13 CN-Fuzzy architecture 80
5.5 Experimental Results
5.5.1 Experimental Mobile Robot Description

This section describes the characteristic of the experimental mobile robot (Figure 5.14).
The robot has two front wheels, which is powered by separate DC geared motors. The
motor driver is used to control the velocity and direction of the robot. The width of the
77
robot plate is 23cm, and the track width and height of robot are 30cm and 8cm,
respectively. The mobile robot is equipped with one sharp infrared range sensor on the
front side, and the two ultrasonic range finder sensors fitted on the left and right side of
the robot, as shown in Figure 5.15. Each sensor can read obstacle from 20cm to 150cm
approximately. The minimum and maximum velocities of the experimental mobile robot
are between 6.7 cm/sec to 16.7 cm/sec approximately.
5.5.2 Experiments
This section presents the experimental results of a mobile robot using CN-Fuzzy
architecture in the different environments. The experiments have been performed by
C/C++ running Arduino microcontroller based mobile robot. The proposed architecture
controls the motor velocities (right and left) of the robot during navigation in the
environment using sensor data interpretation. Figures 5.16 to 5.18 show the real time
navigation of the experimental mobile robot in the different environments. The width and
height of the platform are 250cm and 250cm, respectively. In the experimental results, it
is assumed that the position of the start point and goal point are known. But the positions
of all the obstacles in the environment are unknown for the robot. Firstly, the robot goes
towards the goal in the environment, and if the sensor detects the obstacle in the threshold
range, then the proposed architecture controls the velocity of the mobile robot. The
experimental results in the different snapshots verify the effectiveness of the proposed
architecture. Table 5.6 shows the real-time navigation path length and time taken by the
robot in the various unknown environments. Table 5.7 and Table 5.8 illustrates the
travelling path length and navigation time comparison between the simulation and
experimental results, respectively. In the comparison study between the simulation and
78
Figure 5.14: Experimental mobile robot.
Figure 5.15: Sensor distribution of the experimental mobile robot.
79
Y-AXIS
X-AXIS
80
Y-AXIS
X-AXIS
81
Y-AXIS
X-AXIS
82
Table 5.6: Experimental results of a mobile robot navigation in the different

length (cm) (sec)
Figure 5.16 Without obstacle 109 12.4
results
Error between
Figures 5.6 and 5.16 103 109 5.5%
Figures 5.7 and 5.17 89 94 5.32%
Figures 5.13 and 5.18 80 85 5.88%
Error between
Figures 5.6 and 5.16 11.6 12.4 6.45%
Figures 5.7 and 5.17 10.1 10.8 6.48%
Figures 5.13 and 5.18 9.4 10.1 6.93%
83
5.6 Summary
In this chapter, the CN-Fuzzy architecture has been applied to the intelligent navigation
of a mobile robot in unknown environments filled with obstacles. The major
contributions of this present chapter are summarized as follows: -
1) The cascade neural network is designed to train the robot to reach the goal in the
environment. The inputs of cascade neural network are the obstacle distances, and
the output is the turning angle between the robot and goal.
2) The fuzzy logic controller helps the robot to control the right motor velocity and
left motor velocity in the environments for obstacle avoidance.
3) The proposed CN-Fuzzy architecture gives better results (in terms of path length)
as compared to previous models [28] and [11], which proves the authenticity of
4) Moreover, the simulation and experimental results in the different environment
show the effectiveness of the proposed architecture in the both static and dynamic
environments. In the comparison study between the simulation and experiment
results errors are recorded, and the errors are found due to the effect of slippage
and friction between the wheels of the robot and surface during navigation in real
time mode.
5) During experiment utmost care has been taken to minimize the slippage and
friction between the wheels and surface. Still the effect of slippage and friction are
unavoidable, and errors are recorded during the comparison of the results for
travelling path length (5.57%) and for navigation time (6.62%).
84
Chapter 6
Mobile Robot Navigation in Different

Environments using Takagi-Sugeno
Fuzzy Controller and Simulated
Annealing Algorithm Controller
6.1 Introduction
In this chapter, a Takagi-Sugeno fuzzy model with simulated annealing hybrid algorithm
(Fuzzy-SA) has been designed and implemented for the mobile robot navigation and
obstacle avoidance in the different environments. The simulated annealing algorithm is
used to optimize the output value of the fuzzy controller. The ultrasonic range finder
sensor and sharp infrared range sensor are used to calculate the different obstacle
distances, such as front, right, and left obstacle distance for selecting the suitable steering
angle control command in the environment. The objective function for the simulated
annealing algorithm is considered based on the shortest path using the fuzzy model. The
simulation and experimental results show the proposed method is feasible and valid for a
wheeled mobile robot moving in the different environments.
Navigation can be defined as the process of directing the safe movement of a mobile
robot from one point to another with the help of different types of sensors for obstacle
detection in the different environments like indoor, outdoor, and cluttered by using the
various artificial intelligence navigation techniques. The fuzzy logic plays an important
role in dealing with uncertainty when decision making required for mobile robotic
applications for intelligent control like steering control and motor control, etc. Path
planning is a major topic in the field of a mobile robot, which means to figure out an
optimal or near-optimal non-collision path from a start point to a goal point in an
85
environment with the presence of different shape obstacles. In the present work, the
simulated annealing algorithm has been integrated with the Takagi-Sugeno fuzzy
controller to explore the collision-free path planning for a mobile robot. Moreover, the
effectiveness and efficiency of the proposed hybrid method are demonstrated through
MATLAB software simulation and also done in real-time experiments in various
environments.
The application of mixed soft computing techniques such as Neural Network [11],
Fuzzy Logic [28], Simulated Annealing Algorithm [108], Genetic Algorithm [165], Ant
Colony Optimization Algorithm [144], Particle Swarm Optimization [18], Bee Algorithm
[166], Cuckoo Search Algorithm [167] and other nature-inspired algorithms have been
successfully applied by the various researchers in the field of mobile robotics. The
concept of simulated annealing algorithm has come from statistical mechanics [104]. The
simulated annealing is an iterative search algorithm inspired by the annealing of metals
[105]. In the current study, the authors have attempted to solve the path optimization
problem using a fuzzy controller with simulated annealing hybrid algorithm. The purpose
of the navigation problem for the robot is to search an optimal path between the start
point to goal point [95]. Simulated annealing is a stochastic optimization algorithm, and it
is used for finding the global optimization value (either minimum or maximum) by the
objective function [107]. Simulated annealing algorithm is suitable for the local minima
problem avoidance [168].
In the present study, the application of the fuzzy model and simulated annealing
mixed optimization algorithm has been applied to the mobile robot path planning
problem. The contributions of this chapter are organized as follows: Section 6.1 presents
the introduction. The design of Sugeno-type fuzzy logic model is presented in Section
6.2. Optimizing the fuzzy output using a simulated annealing algorithm for mobile robot
navigation is proposed in Section 6.3. Section 6.4 and 6.5 presents the simulation results
and discussion and comparison with previous works to the proposed hybrid controller,
respectively. Section 6.6 represents the experimental setup and its results and discussion
for validating the proposed hybrid controller. Finally, Section 6.7 depicts the summary.
86
6.2 Design of Sugeno-Type Fuzzy Logic Controller
In this section, the fuzzy controller is used to solve the navigation problem of a mobile
robot in the presence of different shape obstacles. This fuzzy logic controller is activated
when the robot detects any obstacle. The proposed membership functions and the fuzzy
rule-based system are used to navigate the robot among the obstacles in any environment.
Fuzzy logic is a method of formulating the mapping relationship between the inputs to an
output using the mathematical concept. The Fuzzy Inference Systems (FIS) have been
classified into two groups: Mamdani-type and Takagi-Sugeno type. The proposed fuzzy
controller is used to control the steering angle of the mobile robot, which allows the robot
to move from start point to goal point in the environment. The fuzzy controller has three
input variables: x1 (Front Obstacle Distance), x2 (Right Obstacle Distance), and x3 (Left
Obstacle Distance) and single output: f (Steering Angle). The proposed controller
receives inputs (obstacle distances) from the various sensors of the robot. Two
Generalized Bell-Shaped (gbell) membership functions (MFs) are considered for the
inputs. The range of inputs is divided into two linguistic variables: near and away
respectively. The two constant type membership functions (MFs) negative and positive
respectively have been used for the output, and it is located at -90 and 90 respectively.
Figures 6.2 and 6.3 show the membership parameters of the input and output variables
respectively. The fuzzy controller is composed through zero-order Takagi-Sugeno model
in the following form: -
Rulei : If x1 is Ai1 , x2 is Ai 2 , And x3 is Ai 3 THEN fi is i (6.1)
where the symbols x1 , x2 , and x3 are the input variables, Ai1 , Ai 2 , and Ai 3 are the fuzzy
sets, and  i is a real number. The fuzzy set Aij uses the following generalized bell-
shaped membership function: -
ij  x j  
1
2 bij
(6.2)
x j - cij
1
aij
87
where cij is the center of the generalized bell-shaped membership function, aij is the half
width of the membership function, and bij (together with aij ) is controlling the slopes at
the crossover points. The general structure of the generalized bell-shaped membership
function is shown in Figure 6.1. The firing strength  i ( x) is calculated by the following
function: -
n
 i ( x)   ij ( x j ) (6.3)
j 1
The defuzzification of the output (steering angle) is accomplished by the weighted

average method: -
  ( x)  
8
 n 1 i i
fi (6.4)
  ( x)
8
n 1 i
Values
Figure 6.1: Generalized bell-shaped membership function.
88
To avoid the obstacles, the controller drives with user defined rule system based on
human experience, and its functioning under Takagi-Sugeno type fuzzy controller is
shown in Figure 6.4. Eight fuzzy control rules have been designed for the proposed
navigation controller; and are listed in Table 6.1. From the table analysis (fuzzy rule
number three) when the mobile robot comes near the left obstacle, then the robot will turn
to the right side (negative steering angle) for reaching the goal without any collision.
From the fuzzy rule number four, when the mobile robot comes near the right obstacle,
then the robot will turn to the left side (positive steering angle) for reaching the goal
without any collision. The negative steering angle control command means the right
motor velocity will be low, and the left motor velocity will be high respectively. The
positive steering angle control command means the right motor velocity will be high, and
the left motor velocity will be low respectively.
The rule viewer of the proposed fuzzy controller is shown in Figure 6.5, which
represents the eight fuzzy rules in the Graphical User Interface (GUI) form. The first
three input column corresponds to the Front Obstacle Distance (F.O.D.), Right Obstacle
Distance (R.O.D.), and Left Obstacle Distance (L.O.D.) respectively. These obstacle
distances have received from the sensors. The fourth column represents the Steering
Angle (S.A.), which is calculated by the weighted average method (equation (6.4)) of the
fuzzy controller. The steering angle surface plot generated by the fuzzy controller is
shown in Figure 6.6. The proposed fuzzy model has been generated through the
MATLAB programming, which is described below (MATLAB Fuzzy Logic Function).
89
MATLAB FUZZY LOGIC FUNCTION
Program:
1: begin:
2: fuzzy inference system
3: fuzzification
4: add variable for input1 (F.O.D.)
5: add generalized bell-shaped (gbellmf) membership function for input1 (F.O.D.)
6: add variable for input2 (R.O.D.)
7: add generalized bell-shaped (gbellmf) membership function for input2 (R.O.D.)
8: add variable for input3 (L.O.D.)
9: add generalized bell-shaped (gbellmf) membership function for input3 (L.O.D.)
10: add variable for output (S.A.)
11: add Sugeno-type constant membership function for output (S.A.)
12: rule base
13: add rules
14: defuzzification
15: result crisp output (S.A.)
16: end
90
Table 6.1: Fuzzy control rules for mobile robot navigation using two-membership
functions
Front Obstacle Right Obstacle Left Obstacle Steering

Fuzzy rules Distance Distance Distance Angle
(F.O.D.) (R.O.D.) (L.O.D.) (S.A.)
1 Away Away Away Negative
2 Near Near Near Positive
3 Away Away Near Negative
4 Away Near Away Positive
5 Near Away Away Positive
6 Near Away Near Negative
7 Near Near Away Positive
8 Away Near Near Positive

(cm)
Figure 6.2: Fuzzy membership functions for the inputs (F.O.D., R.O.D., and L.O.D.).
91
Figure 6.3: Fuzzy membership function Sugeno-Type for output variable steering angle.
Figure 6.4: Takagi-Sugeno type fuzzy controller.
92
Figure 6.5: Rule viewer of the fuzzy controller.
(degree)
(cm)
(cm)
Figure 6.6: Steering angle control surface function plot.
93
6.3 Optimizing the Fuzzy Controller Output using
Simulated Annealing Algorithm (SAA)
Simulated Annealing Algorithm (SAA) is an iterative search metaheuristic algorithm,

which is inspired by the annealing of metals. SAA is a stochastic optimization algorithm,
and is used for finding the global optimization value (either minimum or maximum) by
the objective function. The SAA starts from an initial solution at a high temperature and
makes some changes according to annealing schedules. For any two iterations, there are
two objective values marked as f i (new) and fi 1 (old), and the difference between the
objective values (f  fi  fi 1 ) is calculated. If f  0 , then the new solution is accepted
with probability p(T )  1 . Otherwise, it is accepted with small probability p(T ) ,
 f 
p(T )  exp   , where k is the constant (Boltzmann) parameter of the process and T
 k T 
is the instantaneous temperature. After enough numbers of iteration, we get the optimum
value of the solution. During SAA process, the objective is move from the high energy
region to low energy region in the search space.
In this section, the result of fuzzy output (consequent) has been optimized through the
simulated annealing algorithm. The fuzzy controller is used to receive the obstacle
distance from the group of sensors to control the steering angle of the wheeled mobile
robot. And the simulated annealing algorithm is applied to find the optimum steering
angle of the mobile robot by using an objective function. The parameters C1 , C2 , C3 , and
C4 are optimized through SAA to acquire optimum steering angle for mobile robot
navigation using objective function. This optimum steering angle is helping to achieve
the minimum path length in the environment. The following objective function calculates
the optimal steering angle (O.S.A.) for a mobile robot: -
O.S. A.  C1  F.O.D.  C2  R.O.D.  C3  L.O.D.  C4  S.A. (6.5)
where,
20  F.O.D.  150
94
20  R.O.D.  150
20  L.O.D.  150
90  S. A.  90
Here the fuzzy rule-based controller has been integrated with the SAA controller to
optimize the steering angle of the mobile robot in different environments. The resulting
Fuzzy-SA controller for navigation of the mobile robot is shown in Figure 6.7. The role
of the fuzzy controller is to estimate the initial steering angle for the hybrid controller.
Then, this initial steering angle is fed to an SAA controller along with the obstacle
distance (front, right and left), and finally this hybrid controller does optimize the steering
angle using a fitness function (equation (6.5)). Figure 6.8 shows the objective function
value versus the number of iteration number. The simulated annealing algorithm model
has been made by MATLAB function programming, which is listed below (MATLAB
Simulated Annealing Algorithm (SAA) Function).
MATLAB SIMULATED ANNEALING ALGORITHM (SAA) FUNCTION

Program:
1: function:
2: Y = new_simulated_annealing(X, F.O.D., R.O.D., L.O.D., S.A.)
3: C1 ← X(1);
4: C2 ← X(2);
5: C3 ← X(3);
6: C4 ← X(4);
7: Y = X (1)*F.O.D. + X (2)*R.O.D. + X (3)*L.O.D. + X (4)*S.A.;
8: end function
95
Figure 6.7: Fuzzy-SA controller for navigation of a mobile robot.
Figure 6.8: Objective function value versus number of iteration number.
96
6.4 Simulation Results and Discussion
In this section, simulations are presented to demonstrate the effectiveness of the singleton
fuzzy controller and Fuzzy-SA controller. The simulation results are performed by using
MATLAB software. The developed architecture of mobile robot navigation based on
Fuzzy-SA algorithm is given in Figure 6.9. Figures 6.10 to 6.12 show the mobile robot
navigation and obstacle avoidance in the different environments using singleton fuzzy
controller and Fuzzy-SA controller. Similarly, the Figure 6.13 demonstrates the
navigation of a mobile robot in an unknown environment with the presence of two
dynamic obstacles using Fuzzy-SA controller. It is assumed that the position of the start
point and goal point are known. But the positions of all the obstacles in the environment
are unknown for the robot. The red line indicates the mobile robot navigation using
singleton fuzzy controller and the green line indicates the mobile robot navigation using
Fuzzy-SA controller. The following criteria are used to evaluate the performance
comparison between the singleton fuzzy controller and the Fuzzy-SA controller: -
(1) Navigation path length (from the start point to goal point).
(2) Navigation time (from the start point to goal point).
The comparison of results between the singleton fuzzy controller and the Fuzzy-SA
controller has been listed in Table 6.2. From the table, it can be seen that the proposed
Fuzzy-SA controller gives better results compared to the singleton fuzzy controller.
97
Begin
Initialize the Start and Goal Position of Mobile

Robot
Mobile Robot Proceed Towards the Goal
YES
Reach Goal Quit
NO
NO On-board Sensors are

Detected any Obstacles
YES
Fuzzy Rule Controller Active and Provides the Initial Steering Angle for
Obstacle Avoidance
Use of Simulated Annealing Algorithm for Optimization of Fuzzy Fired

Steering Angle
YES NO
Obstacles Avoided
Figure 6.9: The developed architecture of mobile robot navigation based on Fuzzy-SA
algorithm.
98
(a)
(cm)
(cm)
(b)
(cm)
(cm)
Figure 6.10: Mobile robot navigation among the single obstacle using (a) Fuzzy
controller and (b) Fuzzy-SA controller.
99
(a)
(cm)
(cm)
(b)
(cm)
(cm)
Figure 6.11: Mobile robot navigation among the many obstacles using (a) Fuzzy
100
(a)
(cm)
(cm)
(b)
(cm)
(cm)
Figure 6.12: Mobile robot navigation among the polygonal obstacles using (a) Fuzzy
101
(cm)
(cm)
(cm) (cm)
(cm)
(cm)
(cm) (cm)
Figure 6.13: Mobile robot navigation in the dynamic environment using Fuzzy-SA
controller.
102
Table 6.2: The result comparison between the singleton fuzzy controller and the Fuzzy-
SA controller
Percentage Percentage
Navigation
deviation of Navigation deviation of
Figure no. Controller path length
navigation time (sec) navigation
(cm)
path length Time
Figure 6.10 Fuzzy 126 8.4

1.61 2.44
(a), (b) Fuzzy-SA 124 8.2

2.27 4.54
(a), (b) Fuzzy-SA 132 8.8

3.51 5.33
(a), (b) Fuzzy-SA 114 7.5
In this section, a comparison has been made between the proposed Fuzzy-SA controller
and the previous models [107], [37] in the simulation and experimental modes. The
performance of this hybrid model is evaluated on the basis of navigation path length.
Martinez-Alfaro et al. [107] have developed the simulated annealing and fuzzy logic
for generating an automatic path planning of the mobile robot. The simulated annealing
algorithm is used to search a collision-free optimal trajectory between the fixed polygonal
obstacles. Forty-nine fuzzy rules have prepared to adjust the velocity of the robot based
on the sonar readings of the obstacle. Liu et al. [37] have studied the path planning
problem of an autonomous mobile robot based on ultrasonic range finder sensor
information by combining the genetic algorithm with the fuzzy inference system.
Figures 6.14 (a) and 6.14 (b) present the path covered by the robot using the previous
model [107] and the proposed hybrid model respectively, in the same environment.
Similarly, Figures 6.15 (a) and 6.15 (b) illustrate the simulation result of the fuzzy-
genetic model [37] and the proposed hybrid model respectively, in the same environment.
103
Martinez-Alfaro et al. [107] model (Table 6.3) robot finds the goal with the
navigation path length of 266cm, and the proposed hybrid model finds the goal with the
optimized navigation path length of 252cm. In Liu et al. [37] method (Table 6.4) robot
finds the goal with the navigation path length of 44cm, and the proposed hybrid model
finds the goal with the optimized navigation path length of 39cm. From the table analysis
below, it can be clearly seen that the proposed Fuzzy-SA controller gives better results
compared to the previous models. The centimeter measurements are taken on the
proportional basis.
(a)
Y-AXIS (cm)
X-AXIS (cm)
104
(b)
(cm)
(cm)
Figure 6.14: The graphical comparison between the (a) Martinez-Alfaro et al. [107]
model and (b) Proposed hybrid model.
Table 6.3: The simulation result comparison between the Martinez-Alfaro et al. [107]
model and proposed hybrid model
Navigation path length of

Optimized navigation path
Figure no. Martinez-Alfaro et al. [107]
length in (cm)
model (cm)
Figure 6.14 (a), (b) 266 252
105
(a)
Y-AXIS (cm)
X-AXIS (cm)
(b)
(cm)
(cm)
Figure 6.15: The graphical comparison between the (a) Liu et al. [37] model and (b)
Proposed hybrid model.
106
Table 6.4: The result comparison between the Liu et al. [37] model and proposed hybrid
model
Navigation path length of Liu et Optimized navigation path

Figure no.
al. [37] model (cm) length (cm)
Figure 6.15 (a), (b) 44 39
6.6 Experimental Results and Discussion
6.6.1 Mobile Robot Description
To demonstrate the effectiveness of the proposed model, real-time experiments have been
conducted using a two-wheeled mobile robot (Figure 6.16) in unknown environments.
The robot has two front wheels driven by two center shaft 12Volt DC geared motors
separately used for steering angle or driving (motor speed) control, and one
omnidirectional caster wheel has been attached on the back side for balance. Two
separate 12Volt DC motors are connected to a dual DC motor controller, and the
microcontroller is used to drive each front wheel to facilitate turn left and right, backward
and forward movements, and the motor voltage regulated by Pulse Width Modulation
(PWM) signal. The width of the robot plate is 23cm, the distance between the wheels is
around 30cm, wheel diameter is 10cm, and width of the wheels is 4cm. The mobile robot
is equipped with one sharp infrared range sensor on the front side, and the two ultrasonic
range finder sensors are equipped on the left and right side of the robot, as shown in
Figure 6.17. Each sensor reads ranges from 20 cm to 4m approximately. The maximum,
and minimum velocity of mobile robot used for navigation is Vmax = 0.167 m/sec, and Vmin
= 0.067 m/sec respectively.
6.6.2 Experiments
In during the experiments, the mobile robot is controlled by intelligent processor. The
width and height of the experimental platform are 350cm and 350cm, respectively. The
environment is assumed to be entirely unknown for the robot, and the sensing range of
107
the on-board robot sensors is limited. The average moving speed of the robot is 0.09
m/sec. The experimental verification of the above simulation results has been shown in
Figure 6.18 and Figure 6.19. In Figure 6.18, the single obstacle is used for the navigation.
Similarly, in Figure 6.19, three obstacles are used for the navigation. It is assumed that
the position of the start point and goal point are known, but the positions of all the
obstacles in the environment are unknown for the robot. From the experimental analysis,
it can be clearly seen that the proposed Fuzzy-SA controller successfully avoid the
obstacles in the given environment. The navigation path length result between the
simulation and experimental is listed in Table 6.5. Tables 6.6 and 6.7 illustrate the
experimental results. In the comparison study between the simulation and experiments, it
is observed that some errors have been found, these are happened due to slippage and
friction during real time experiment.
Figure 6.16: Two-wheeled mobile robot.
108
Figure 6.17: The arrangement of the sensors of a mobile robot.
109
Y-AXIS
X-AXIS
110
Y-AXIS
X-AXIS
111
Table 6.5: Navigation path lengths between simulation and experimental results
Navigation Path Length (cm)

Figure no. (Simulation
and experimental res.)
Simulation Result Experimental Result
Figure 6.10 (b) and 6.18 124 131
Figure 6.15 (b) and 6.19 39 42
results
Error between
Figures 6.10 (b) and 6.18 124 131 5.34%
Figures 6.15 (b) and 6.19 39 42 7.14%
Error between
Figures 6.10 (b) and 6.18 8.2 8.8 6.82%
Figures 6.15 (b) and 6.19 4.6 4.9 6.12%
112
6.7 Summary
In the current study, a Takagi-Sugeno fuzzy model and simulated annealing hybrid
algorithm (Fuzzy-SA) have been proposed for the mobile robot navigation and obstacle
avoidance in the different environments (static and dynamic). The role of the fuzzy
controller is to estimate the initial steering angle of the mobile robot, and the simulated
annealing algorithm is applied to optimize this initial steering angle using objective
function (see the equation (6.5)). Effectiveness and efficiency of this proposed Fuzzy-SA
controller have been verified through many simulation and experimental studies in the
real environment. In the comparison study between the simulation and experiment results
errors are recorded, and the errors are found due to the effect of slippage and friction
between the wheels of the robot and surface during navigation in real time mode. During
experiment utmost care has been taken to minimize the slippage and friction between the
wheels and surface. Still the effect of slippage and friction are unavoidable, and errors are
recorded during the comparison of the results for travelling path length (6.24%) and for
navigation time (6.47%). Moreover, this Fuzzy-SA controller is compared with previous
Fuzzy-Simulated Annealing Algorithm [107] and the Fuzzy-Genetic Algorithm [37] to
prove the authenticity of the proposed controller.
113
Chapter 7
Optimum Navigation of a Mobile Robot

in the Different Environments using
Wind Driven Optimization Algorithm
7.1 Introduction
In this chapter, atmospheric motion based Wind Driven Optimization (WDO) algorithm
is implemented to optimize the navigation path length of a mobile robot in the various
environments with the presence of different shape obstacles. This optimization algorithm
is working based on the atmospheric motion of infinitesimal small air parcels navigates
over an N-dimensional search domain. The WDO will be used to find the optimum or
near-optimum steering angle for the navigation of a mobile robot to achieve the minimum
path length in given environment. The consideration of the objective function is depended
mainly for two reasons, i.e. obstacle avoidance and to find out the shortest possible path
in an environment. Simulation and experimental results demonstrate that the robot can
determine the optimum path length using Wind Driven Optimization (WDO) compared to
the path length obtained by the robot using Genetic Algorithm (GA) and Particle Swarm
Optimization (PSO) algorithm.
Metaheuristic optimization methods such as Genetic Algorithm [169], Simulated
Annealing Algorithm [107], Particle Swarm Optimization [121], Ant Colony
Optimization [148], Cuckoo Search [167], Firefly Algorithm [170], and many other
nature-inspired optimization algorithms have been proposed and successfully
implemented for the robot path planning problems in static as well as dynamic
environments, but each algorithm possesses strengths and weakness. In the recent years,
there are many researchers focused on meta-heuristic optimization in the field of mobile
robot path optimization problems. Other soft computing techniques like Fuzzy logic
114
[171], Neural Network [80], ANFIS [172], and its hybrid methods [107,15] are also
applied successfully in the various research area of mobile robot path planning in
different environments. In this chapter, new type atmospheric motion based, a nature-
inspired global optimization technique is implemented for mobile robot navigation in the
unknown static and dynamic environments. The Wind Driven Optimization [153-154]
method is a population-based iterative heuristic optimization algorithm for multi-
dimensional and multi-model problems with the potential to implement constraints on the
search domain [156]. As a newly intelligent optimization algorithm, WDO algorithm has
been successfully applied in the different optimization fields [155-157] and achieved
better results in some cases.
Path planning for a mobile robot finds an optimum and safe path from start point to a
target point in an environment with obstacles. In this chapter, a new nature-inspired
atmospheric motion based WDO algorithm is applied to solve the path planning problem
of the mobile robot in unknown cluttered environments. WDO algorithm works by
simultaneously maintaining several infinitesimal small air parcels or potential solutions in
the search domain. For each iteration of the algorithm, each air parcels are evaluated by
the objective function being optimized based on the fitness function of that solution.
Consideration of the objective function is based on two reasons, i.e. obstacle avoidance
and find out the shortest possible path in an environment. In the present study, an attempt
is made to verify the effectiveness of the WDO algorithm based optimized navigation of a
robot in unknown cluttered and dynamic environments filled with different shape
obstacles. This optimization algorithm has been successfully implemented in the different
areas of the engineering applications [155-157]. Due to its broad area of application and
performance, this optimization algorithm has been used for solving the mobile robot’s
optimum path planning problem. In order to demonstrate the success of this new
optimization technique, it is applied to different unknown simulation and experimental
conditions and compared with previous techniques like adaptive Genetic Algorithm
[169], Particle Swarm Optimization [173] and found to be good agreement in terms of
navigation path length.
This chapter is structured as follows: Section 7.1 presents the introduction. Path
optimization using Wind Driven Optimization (WDO) algorithm for mobile robot
navigation is proposed in Section 7.2. Section 7.3 presents the computer simulation
115
results and discussion. Comparison with the previous navigational controller with the
proposed controller is listed in Section 7.4. Section 7.5 represents the experimental setup
and its results. Finally, Section 7.6 depicts the summary.
7.2 Path optimization using Wind Driven Optimization

(WDO) algorithm
Wind driven optimization [153] algorithm is inspired from the earth’s atmosphere, where
the blows of wind are trying to equalize the horizontal imbalance in the air pressure.
WDO is a new type nature-inspired global optimization based on atmospheric motion
developed by Bayraktar et al. [153] in 2013. This method is working on the population-
based iterative heuristic global optimization algorithm for multi-dimensional and multi-
modal problems with the potential to implement constraints on the search domain. WDO
is similar to other nature-inspired optimization algorithms, in which population-based
heuristic iterative process can be found for solving multi-dimensional optimization
problems [157]. At its center, a population of infinitesimally small air parcels navigates
over an N-dimensional search space, employing Newton’s second law of motion that is
used to express the motion of air parcels inside the earth’s atmosphere. As compared to
similar particle based optimization algorithm (e.g., PSO), the WDO algorithm has
additional terms in the velocity update equation such as Gravitation and Coriolis forces,
which provides robustness and extra degrees of freedom to the algorithm.
Along with the theory and terminology of WDO, a numerical study for tuning the
WDO parameters has been introduced in Bayraktar et al. [153]. WDO is further applied
to different applications of electromagnetic optimization problems [153]. These papers
signify that WDO can, in some cases, gives better performance in comparison to other
well-known techniques like PSO, GA, and other optimization algorithms. The WDO
algorithm is working based on the atmospheric motion of infinitesimal small air parcels
navigates over an N-dimensional search domain. The starting step of this algorithm is
supported by the Newton’s second law of motion, which provides accurate results when
applied to the analysis of atmospheric motion. It states that the total force applied on an
air parcel causes it to accelerate with an acceleration a in the same direction as the
applied total force.
116
  a   Fi (7.1)
where  is the density of air for an infinitesimally small air parcel, and Fi represents all
the individual forces acting on the air parcel. To relate the air pressure to the air parcel’s
density and temperature, the ideal gas law can be utilized and is given by: -
P   RT (7.2)
where P is the pressure, R is the universal gas constant, and T is the temperature.
Four major forces can be included in the equation (7.1) that either cause the wind to
move in a certain direction at a certain velocity or that deflect it from its existing path.
The most observable force causing the air to move is the Pressure Gradient force FPG
defined in equation (7.3). Another force is the Friction force FF defined in equation (7.4),
which simply acts to oppose the motion started by the Pressure Gradient force. In the
three-dimensional physical atmosphere, the Gravitational force FG in equation (7.5) is a
vertical force directed toward the earth’s surface. The Coriolis force FC in equation (7.6)
is caused by due to the rotation of the earth and deflects the path of wind from one
dimension to another.
FPG  P  V (7.3)
FF       u (7.4)
FG     V  g (7.5)
FC  2  u (7.6)
where, P is the pressure gradient,  V represents the infinite air volume,  represents
the rotation of the earth, g is the gravitational acceleration,  is the friction coefficient
and u is the velocity vector of the wind.
The sum of all forces (Gravitational force, Pressure Gradient force, Friction force, and
Coriolis force) described above can be entered on the right-hand side of Newton’s second
law of motion given in equation (7.1), which leads to: -
117
u
      V  g    P   V         u    2   u  (7.7)
t
where the acceleration term in equation (7.1) is rewritten as a  u t , and a time step
t  1 is assumed for simplicity. For an infinitesimally small, dimensionless air parcel,
the volume is set as  V  1, which simplifies the equation (7.7) to: -
  u     g    P         u    2  u  (7.8)
Putting the ideal gas law equation (7.2) in equation (7.8), the density  can be written in
terms of the pressure, with temperature and the universal gas law constant: -
   c  ucur
other dim

unew  1     ucur  g  xcur   RT  1  xopt  xcur    
1
 (7.9)
 i   i 
where unew is the velocity in the next iteration, ucur is the velocity in current iteration, xcur
is the current location of the air parcel, xopt is the optimum location of the air parcel, i
other dim
represents the ranking between all air parcels, ucur is the velocity influence from
another randomly chosen dimension of the same air parcel, and all other coefficients are
combined into a single term c (i.e., c  2  RT ). Equation (7.9) represents the final
form of the velocity update utilized in WDO [153, 155]. The following function updates
the position of the air parcel: -
xnew  xcur  (unew  t ) (7.10)
where xnew is the new position of the air parcel in the next iteration. If the new velocity
unew exceeds the initialize maximum velocity ( umax = 0.3) in any dimension, then the
velocity in that dimension is limited according to the following condition: -
  u if unew  umax
unew   max (7.11)
umax if unew  umax
where the direction of motion is preserved but the magnitude is limited to be no more

than umax at any dimension and unew represents the adjusted velocity after it is limited to
the maximum speed.
118
The pseudo-code of the WDO algorithm can be summarized as follows: -
Step 1. Start.
Step 2. Initialize the population size (i.e., group of air parcels), number of dimensions of
the optimization problem, maximum number of iterations, coefficients (such as RT , g ,
 , c , umax ), pressure function (fitness function of the optimization problem), lower and
upper boundaries of the optimization problem.
Step 3. Assign random position and velocity of the air parcels.
Step 4. Evaluate the pressure (fitness) values of each air parcel at its current position.
Step 5. Once the pressure values have been evaluated, the population is ranked based on
their pressure (ascending order), and the velocity updated according to equation (7.9)
along with the restrictions are given in equation (7.11).
Step 6. Update the position of the air parcel for the next iteration according to equation
(7.10) and also check the boundaries of the air parcel.
Step 7. Stop if a maximum number of iterations are achieved, else go to step 4.
When the maximum number of iterations is completed, the best pressure (objective)
value is achieved.
In the Wind driven optimization (WDO) based path planning problems, the objective
function is considered as the best steering angle relies on the criterion of the random
search model. This optimum steering angle is helping to achieve the minimum path
length in the given environment. The objective function measures the optimum solution
of the current model. In this chapter, the objective function values are Front Obstacle
Distance (F.O.D.), Right Obstacle Distance (R.O.D.), and Left Obstacle Distance
(L.O.D.). Obstacle distance is used to optimize the following objective equation (7.12)
for getting optimum Steering Angle (S.A.) for the proposed model:-
S.A.  11.217  0.0016  F.O.D.  0.7326  R.O.D.  1.1317  L.O.D. (7.12)
where,
20  F.O.D.  150 (Ranging between 20 to 150 same as sensor data range)
20  R.O.D.  150 (Ranging between 20 to 150 same as sensor data range)
20  L.O.D.  150 (Ranging between 20 to 150 same as sensor data range)
119
The selection of the objective function is based on two reasons, i.e. obstacle
avoidance and find out the shortest possible path in an environment. These obstacle
distances (F.O.D., R.O.D., and L.O.D.) are acquired by the different equipped sensors
such as ultrasonic range finder sensor, sharp infrared range sensor, and other sensors. The
sensing range of sharp infrared range sensor is between 20cm to 150cm, and the
ultrasonic range finder sensor is 2cm to 4m. Both sensors are used for sensing the front,
left, and right placed obstacles around the robot. If the robot moves in an environment
filled with static and dynamic obstacles, it is necessary to detect the obstacles and avoid
it. The developed architecture of mobile robot navigation based on WDO method is given
in Figure 7.1.
7.3 Computer Simulation Results and Discussion
In order to verify the effectiveness of the proposed WDO algorithm, the different
simulations have been done using MATLAB software. The WDO algorithm ran with the
different population size of 20, 30, 40 air parcels, and three number of dimensions for a
maximum of 500 epochs (iterations). Other necessary parameters used for WDO, GA,
and PSO algorithm are listed in Tables 7.1 to 7.3, respectively. Simulation results in
Figures 7.2 to 7.6 demonstrate the mobile robot navigation and obstacle avoidance in the
different static and dynamic environments. It is assumed that the position of the start
point and goal point are known. But the positions of all the obstacles in the environment
are unknown for the robot. The obstacle distance information obtained by the equipped
sensors are the inputs of the navigational controller, which provides the necessary
optimum steering angle control command as its output for navigation. In order to
demonstrate the efficiency of this new optimization technique, it is applied to different
unknown simulation environments and compared with other optimization techniques like
Genetic Algorithm, Particle Swarm Optimization and the results are found to be in
agreement in terms of navigation path length and also the time taken to reach the target
(Tables 7.4 and 7.5) and are illustrated in Figures 7.7 to 7.8. The average deviation path
length between WDO versus GA is 0.7%, and the WDO versus PSO is 1.640%.
120
Begin
Mobile Robot Proceeds Towards the Target
Reach Target YES Quit
NO
NO Robot Near to Obstacles

on the Target Path
YES
WDO Algorithm Active
Initialization: Population Size, No. of Iterations

Coefficients, Boundaries, and Objective Function
Calculate the Objective Function Value

Equation (7.12)
Calculate the Best Global Minimum Pressure
Robot Moves According to Given Global Minimum

Pressure (Steering Angle)
YES NO
Obstacles Avoided
Figure 7.1: The architecture of mobile robot navigation based on WDO method.
121
Table 7.1: Parameters used in WDO algorithm
S. No. Parameters Values
1. Population size 20
2. Dimension of the problem 3
3. Maximum number of iterations 500
4. RT coefficient 3
5. Gravitational constant 0.2
6. Friction coefficient 0.4
7. Coriolis effect 0.4
8. Velocity limit 0.3
Table 7.2: Parameters used in GA
Sl. No. Parameters Values/Function Name
1. Population size 20
2. Selection function Stochastic uniform
3. Elite count 1.8
4. Crossover fraction 0.7
5. Mutation function Constraint dependent
6. Crossover function Scattered
7. Number of generation 100
122
Table 7.3: Parameters used in PSO algorithm
Sl. No. Parameters Values
1. Swarm size 150
2. Maximum number of iterations 400
3. Social acceleration factors 2
4. Cognitive acceleration factors 2

(cm)
(cm)
Figure 7.2: Navigation of a mobile robot using WDO technique.
123
(cm)
(cm)
Figure 7.3: Navigation of a mobile robot using GA technique.

(cm)
(cm)
Figure 7.4: Navigation of a mobile robot using PSO technique.
124
(cm)
(cm)
(cm) (cm)
(cm)
(cm)
(cm) (cm)
Figure 7.5: Navigation of a mobile robot using WDO technique in a dynamic

environment.
125
(cm)
(cm)
Figure 7.6: Navigation of a mobile robot using WDO technique in a cluttered

environment.
Figure 7.7: Comparison performance graph between WDO algorithm over GA, and PSO
in terms of navigation path length.
126
Figure 7.8: Comparison performance graph between WDO algorithm over GA, and PSO
in terms of time taken to reach the target.
Table 7.4: Comparison the performance of WDO algorithm over GA and PSO in terms of
navigation path length
Navigation path length Navigation path length Navigation path length

(cm) of the a robot using (cm) of the a robot using (cm) of the a robot using
WDO in Figure 7.2 GA in Figure 7.3 PSO in Figure 7.4
98 100 103
97 101 102
99 101 103
100 103 105
98 104 101
Note: Bold value indicates the minimum path length.
127
Table 7.5: Comparison the performance of WDO algorithm over GA and PSO in terms of
time taken to reach the target
Time taken (sec) to reach Time taken (sec) to reach Time taken (sec) to reach
the target by the robot the target by the robot the target by the robot
using WDO in Figure 7.2 using GA in Figure 7.3 using PSO in Figure 7.4
8.92 9.16 9.33
8.9 9.19 9.35
9.01 9.21 9.44
9.11 9.4 9.55
8.97 9.41 9.21
Note: Bold value indicates the minimum time.
7.4 Comparison with Previous Navigational Controllers
A comparison has been done on path generated by the Genetic Algorithm [169], and
Particle Swarm Optimization [173] over the current Wind Driven Optimization algorithm
based navigation technique in both simulation and experimental modes. The performance
of this method is evaluated on the basis of navigation path length.
Jianguo et al. [169] have proposed an adaptive Genetic Algorithm (GA) method that
automatically adjusts the GA parameters like crossover and mutation, according to the
different changeable environment for a mobile robot. They also compared this adaptive
Genetic Algorithm method to traditional GA and stated that the adaptive Genetic
Algorithm performs better over traditional GA.
Deepak et al. [173] have presented Particle Swarm Optimization based technique that
avoids the obstacle efficiently and generates a feasible path in unknown environments.
They performed various graphical simulation tests in a different scenario to verify the
effectiveness of the developed technique.
Figures 7.9 (i) and 7.9 (ii) show the comparison between path generated by GA [169]
and the current WDO algorithm for the same start position to target position. WDO
128
generated an optimum path in comparison to GA for the same path planning problem.
Similarly, Figures 7.10 (i) and 7.10 (ii) illustrate the comparison between path generated
by PSO [173] and the current WDO algorithm for the same start position to target
position. WDO gives an optimum path in comparison to PSO for the same path planning
problem.
In Table 7.6, GA [169] based navigational controlled robot finds the target with the
path length of 115 cm and the proposed WDO algorithm based navigational controlled
robot finds the target with an optimum path length of 111 cm. Similarly, in Table 7.7,
PSO [173] based navigational controlled robot finds the target with the path length of 171
cm and the proposed WDO algorithm based navigational controlled robot finds the target
with optimal path length of 160 cm. During comparison (Tables 7.6-7.7), it can be seen
that the proposed controller gives better results (in terms of path length and smoothness)
compared to the previous navigational controller [169, 173]. The centimeter
measurements are taken on the proportional basis.
(i) GA
Y-AXIS (cm)
X-AXIS (cm)
129
(ii) WDO
(cm)
(cm)
Figure 7.9: A simulation comparison results between GA (i) and WDO (ii).
Table 7.6: Results of Jianguo et al. [169] method and WDO algorithm
GA based path length (cm) WDO algorithm based

Figure no.
of Jianguo et al. [169] path length (cm)
Figure 7. 9 (i) and (ii) 115 111
(i) PSO
Y-AXIS (cm)
X-AXIS (cm)
130
(ii) WDO
(cm)
(cm)
Figure 7.10: A simulation comparison results between PSO (i) and WDO (ii).
Table 7.7: The comparison result between Deepak et al. [173] method and WDO
algorithm
PSO based path length WDO algorithm based

Figure no.
(cm) of Deepak et al. [173] path length (cm)
Figure 7.10 (i), (ii) 171 160
7.5 Experimental Results and Discussion
In order to validate the effectiveness of the proposed WDO algorithm, real-time

experiments are done on the real four-wheeled mobile robot (Figure 7.11) in unknown
environments. The main specifications of the proposed prototype experimental mobile
robot are given in Table 7.8. The WDO based navigational controller is implemented in
the Arduino MEGA (ATmega2560) microcontroller. The experimental prototype of four-
wheeled differentially steer controlled mobile robot is shown in Figure 7.11. The motion
and orientation are controlled by independent four DC geared motors, which provide the
necessary torque to all driving wheels. The orientation of the mobile robot is achieved by
131
the differential steered control (motor speed) of all wheels. Four separate 12Volt DC
motors are attached to a dual DC motor driver (L298) and the each driver’s direction and
velocity control pins are connected to the Arduino MEGA microcontroller to drive each
motor to facilitate turn left and right, backward and forward movements, and the motor
velocity regulated by the Pulse Width Modulation (PWM) signal. The mobile robot uses
two kinds of sensors such as two ultrasonic range finder (HC SR-04) sensors and one
sharp infrared range sensor (GP2Y0A02YK0F) shown in Figure 7.12 (Schematic diagram
of differentially steered four-wheeled mobile robot). The sensors are used to prevent a
collision from surrounding obstacles. To sense the obstacles one sharp infrared sensor is
mounted on the front side of the robot for front obstacle detection, and two ultrasonic
range finder sensors are attached to the left and right sides of the robot for left and right
obstacle detection respectively. The two groups of ultrasonic range finder sensors
mounted on the left, and right side of the robot is connected to the digital input port of the
Arduino microcontroller. The single sharp infrared range sensor mounted on the front
side of the robot is connected to the analog input port of the Arduino MEGA
microcontroller.
Experimental verification of the above simulation results has been shown in Figures
7.13 and 7.14. In the real-time experimental result analysis, it can be indicated that the
mobile robot reaches their target position successfully without hitting any obstacles in the
given cluttered environment. It is assumed that the position of the start point and goal
point are known, but the positions of all the obstacles in the environment are unknown for
the robot. These experimental results depict that the mobile robot can find a collision-free
near-optimal path in unknown environments using on-board sensory information. Figure
7.15 illustrates the experimentally obtained navigation paths follow closely those traced
by the robot during simulation results. Table 7.9 shows the real-time navigation path
length and time taken by the robot in the various unknown environments. Tables 7.10 and
7.11 illustrate the travelling path length and navigation time comparison between the
simulation and experimental results. In the comparison study between the simulation and
132
Figure 7.11: Experimental four-wheeled real mobile robot.
Positive Steering Negative Steering

Front I.R.
Angle Angle
Sensor
Left U.S. Right U.S.

Sensor Sensor
Dual DC Motor
Driver
Arduino
Microcontroller
DC Geared
Motors
Figure 7.12: Schematic diagram of differentially steered four-wheeled mobile robot.
133
Table 7.8: Main specifications of the proposed prototype experimental mobile robot
Name Specifications
Microcontroller Arduino MEGA 2560 (ATmega2560).
Flash Memory 256 KB (ATmega2560).
Operating Voltage 5V.
SRAM 8 KB (ATmega2560).
Input Voltage 7–12V (Recommended).
Input Voltage (Limits) 6–20V.
Digital Input/Output
54 (of Which 15 Can be Used as PWM Outputs).
Pins
Analog Input Pins 16.
Motors 4 DC, 30RPM Centre Shaft Economy Series DC Motor.
Motors Driver L298, Up to 46V, 2A Dual DC Two Motor Drivers.
Motor Speed Max: 30RPM, Min: 12RPM.
Wheel Diameter: 106mm, Wheel Thickness: 44mm, Hole

Wheel
Diameter: 8mm.
One Sharp Infrared Range Sensor (GP2Y0A02YK0F)
Distance Measuring Range: 20cm to 150cm.
Sensors
Two Ultrasonic Range Finder Sensor (HC SR-04) Distance
Measuring Range: 2cm to 400cm.
Bread Board Small Size Bread Board.
Communication USB connection Serial Port.
Principal Dimensions Height: 30cm, Length: 62cm, Width: 49cm.
Weight Approx. 4.936kg.
Payload Approx. 500g.
Two Rechargeable Lithium Polymer 3 Cell, 11.1V,

Power
2000mAh, 20C Battery
134
Target
Obstacle
(i) Start (ii)
Target
Y-AXIS
Start
(iii) (iv)
X-AXIS
Figure 7.13: Experimental result of mobile robot navigation same as a simulation

environment (shown in Figure 7.2).
135
Target
Obstacles
(i) Start (ii)

Y-AXIS
(iii) (iv) (v)

X-AXIS
Figure 7.14: Experimental result of mobile robot navigation same as a simulation

environment (shown in Figure 7.9 (ii)).
136
Figure 7.15: Comparison of path length between simulation and experimental results.
Table 7.9: Experimental results of mobile robot navigation in the different environments
using WDO algorithm
Experimental path Navigation time

length (cm) (sec)
Figure 7.13 Single obstacle environment 103 9.5
Figure 7.14 Cluttered environment 117 10.3
137
Table 7.10: Navigation path lengths between simulation and experimental results
Figure no. Navigation path length (cm) Error between

(Simulation and simulation and
Simulation Experimental
experimental res.) experimental result
result result
Figures 7.2 and 7.13 97 103 5.82%
Figures 7.9 (ii) and 7.14 111 117 5.13%
Figure no. Navigation time (sec) Error between

Simulation Experimental
experimental res.) experimental result
result result
Figures 7.2 and 7.13 8.9 9.5 6.32%
Figures 7.9 (ii) and 7.14 9.6 10.3 6.79%
7.6 Summary
In order to demonstrate the success of this new optimization technique, it is applied to the
different unknown cluttered simulation and experimental environments. Comparisons
have been done with previous techniques like adaptive Genetic Algorithm [169], Particle
Swarm Optimization [173] and are found to be in agreement in terms of navigation path
length. The Wind Driven Optimization based path planning for a mobile robot finds an
optimum or near-optimum path from a start point to target point in the static and dynamic
environments filled with obstacles. The simulation and the experimental results show the
proposed optimization algorithm for a mobile robot navigation problem is feasible and
practical. Using this algorithm, the robot can also negotiate with dynamic obstacles
138
successfully. In the comparison study between the simulation and experiment results
errors are recorded, and the errors are found due to the effect of slippage and friction
between the wheels of the robot and surface during navigation in real time mode. During
experiment utmost care has been taken to minimize the slippage and friction between the
wheels and surface. Still the effect of slippage and friction are unavoidable, and errors are
recorded during the comparison of the results for travelling path length (5.48%) and for
navigation time (6.56%). In future, this technique can be hybridized with another
algorithm to improve the navigational performance of the mobile robot.
139
Chapter 8
Optimum Path Planning of Mobile Robot

in Unknown Static and Dynamic
Environments using Fuzzy-Wind Driven
Optimization Algorithm
8.1 Introduction
This chapter introduces a singleton fuzzy (S-Fuzzy) controller and Fuzzy-WDO hybrid
algorithm for the autonomous mobile robot navigation and collision avoidance in an
unknown static and dynamic environment. The WDO (Wind Driven Optimization)
algorithm is used to optimize and tune the input/output membership function parameters
of the fuzzy controller. The WDO algorithm is working based on the atmospheric motion
of infinitesimal small air parcels navigates over an N-dimensional search domain. The
performance of this proposed technique has compared through many computer
simulations and real-time experiments by using Khepera-III mobile robot. As compared
to the S-Fuzzy controller the Fuzzy-WDO algorithm is found good agreement for mobile
robot navigation.
One major problem with the fuzzy logic is the difficulty of constructing and tuning
the correct membership function grade [174]. Therefore, the authors have tried to solve
this problem by using WDO algorithm. In this chapter, a fuzzy-WDO hybrid algorithm
has been presented for mobile robot navigation and collision avoidance in an unknown
static and dynamic environment. The WDO is integrated with the fuzzy controller to
adjust and optimize the antecedent and consequent parameters of the generalized bell-
shaped membership function. The WDO [153] method is a population-based iterative
heuristic global optimization algorithm for multi-dimensional and multi-model problems
140
with the potential to implement constraints on the search domain. This algorithm works
by simultaneously maintaining several infinitesimal small air parcels or potential
solutions in the search domain. For each iteration of the algorithm, each air parcels are
evaluated by the membership function parameters (objective function) being optimized
based on the fitness function of that solution. The primary objective of this research is to
optimize the membership function parameters of the fuzzy controller by using WDO
algorithm.
This chapter is organized into seven sections. Section 8.1 presents the introduction.
Singleton fuzzy (S-Fuzzy) controller for mobile robot navigation is proposed in Section
8.2. The hybrid fuzzy-WDO algorithm for mobile robot navigation is presented in Section
8.3. Section 8.4 demonstrates the simulation results of the mobile robot in different
environments. Section 8.5 describes the simulation result comparison with previous
works. Section 8.6 presents the experimental results and discussion for validating the
proposed controller. Finally, Section 8.7 depicts the summary.
8.2 Singleton Fuzzy (S-Fuzzy) Controller for the Mobile

Robot Navigation
In this section, a singleton fuzzy (S-Fuzzy) rule-based controller has been designed and
implemented for mobile robot navigation and collision avoidance in an unknown static
and dynamic environment. The S-Fuzzy controller is used to control the right motor
velocity and left motor velocity of the mobile robot. The S-Fuzzy controller has three
inputs: Forward Obstacle Distance ( d f ), Left Forward Obstacle Distance ( d l ) and Right
Forward Obstacle Distance ( d r ); and two outputs: Right Motor Velocity ( mr ) and Left
Motor Velocity ( ml ), which are logically connected by eight rules (see the Figure 8.2).
These input and output variables are illustrated in Figures 8.3 and 8.4, respectively. The
fuzzy rule set of the S-Fuzzy controller is described in Table 8.1. The two generalized
bell-shaped (Gbell) membership functions are used for inputs and outputs. The range of
inputs is divided into two linguistic variables: NEAR and FAR. These inputs are located
at 20cm to 150cm. Similarly, the two Gbell membership functions (MFs) LOW and
HIGH, respectively have been used for the outputs, and it is located at 6.7cm/sec to
141
16.7cm/sec. The S-Fuzzy controller is composed through Mamdani-type fuzzy model in
the following form: -
Rulen : IF d f is d f i  , dl is dl  j  , & dr is dr  k  THEN mr is mr ijk  & ml is ml ijk  (8.1)
where n =1, 2,…,8 (eight rules), the i =1, 2, j =1, 2 and k =1, 2 because d f , d l and d r
have two Gbell membership functions each. The d f (i ) , dl ( j ) , and d r ( k ) are the fuzzy sets
of the inputs d f , d l , and d r , respectively. Similarly, the mr ijk  , and ml ijk  are the fuzzy
sets of the outputs mr , and ml , respectively. The fuzzy set (inputs and outputs) uses the
following Gbell membership function: -
Let d f , d l , and d r are presented by x1 , x2 , and x3 respectively. Similarly, mr , and
ml are denoted by y1 , and y2 , respectively.
1
n1  x1   2 bn1
(8.2)
x -c
1  1 n1
an1
1
n 2  x2   2 bn 2
(8.3)
x -c
1  2 n2
an 2
1
n3  x3   2 bn 3
(8.4)
x -c
1  3 n3
an 3
1
n1  y1   2 bn1
(8.5)
y -c
1  1 n1
an1
1
 n 2  y2   2 bn 2
(8.6)
y2 - cn 2
1
an 2
142
where a , b , and c are adjusting parameters of the membership function; called as the
half width, slope control, and center respectively. The general structure of the generalized
bell-shaped membership function is shown in Figure 8.1.
Values
Figure 8.1: The general structure of the generalized bell-shaped membership function.
The defuzzification of the outputs ( y1 and y2 ) are accomplished by the weighted average
method: -
 n1 ( x1 )  n 2 ( x2 )  n 3 ( x3 )   y1
y1  n 1
8 (8.7)
n 1
n1 ( x1 )  n 2 ( x2 )  n 3 ( x3 ) 
 n1 ( x1 )  n 2 ( x2 )  n 3 ( x3 )   y2
y2  n 1
8 (8.8)

n 1
n1 ( x1 )  n 2 ( x2 )  n 3 ( x3 ) 
The adjusting parameters a , b , and c of the inputs and outputs are listed in Tables 8.2
and 8.3, respectively, which will be optimized through the WDO algorithm in Section 8.3
below.
143
Rule 1
Input
[20 150]
Output
Rule 2
[6.7 16.7]
Input
[20 150]
Output
Input [6.7 16.7]
[20 150]
Rule 8
Figure 8.2: The structure of an S-Fuzzy controller for mobile robot navigation.
Table 8.1: Fuzzy rules set
If ( d f is Far) and ( d l is Far) and ( d r is Far) then ( mr is High) and ( ml is Low)
If ( d f is Near) and ( d l is Near) and ( d r is Near) then ( mr is Low) and ( ml is High)
If ( d f is Far) and ( d l is Near) and ( d r is Far) then ( mr is Low) and ( ml is High)
If ( d f is Far) and ( d l is Far) and ( d r is Near) then ( mr is High) and ( ml is Low)
If ( d f is Near) and ( d l is Far) and ( d r is Far) then ( mr is Low) and ( ml is High)
If ( d f is Near) and ( d l is Near) and ( d r is Far) then ( mr is Low) and ( ml is High)
If ( d f is Near) and ( d l is Far) and ( d r is Near) then ( mr is High) and ( ml is Low)
If ( d f is Far) and ( d l is Near) and ( d r is Near) then ( mr is Low) and ( ml is High)
144
Figure 8.3: Fuzzy membership functions for the inputs ( d f , d l , and d r ).

Figure 8.4: Fuzzy membership functions for the outputs ( mr , and ml ).
145
Table 8.2: Adjusting parameters of the inputs before optimization
Membership
Inputs a b c
function
Near 65 2.5 20
df
Far 65 2.5 150
Near 65 2.5 20
dl
Far 65 2.5 150
Near 65 2.5 20
dr
Far 65 2.5 150
Table 8.3: Adjusting parameters of the outputs before optimization
Membership
Outputs a b c
function
Low 5 2.5 6.7

mr
High 5 2.5 16.7
Low 5 2.5 6.7

ml
High 5 2.5 16.7
8.3 Fuzzy-WDO Algorithm for the Mobile Robot

Navigation
This section describes the WDO algorithm used for the membership function parameter
optimization of the S-Fuzzy controller for the optimum navigation and collision
avoidance in an unknown static and dynamic environment. One major problem with the
fuzzy logic is the difficulty of constructing and tuning the correct membership function
grade [174]. Because of this problem, the WDO algorithm is used to tune the adjusting
parameters of the inputs and outputs. From the Section 8.2, two Gbell membership
146
functions are considered for the inputs ( d f , d l , and d r ) and outputs ( mr , and ml ). Each
Gbell membership function has three adjusting parameters ( a , b , and c ). Therefore,

each input has six adjusting parameters. Similarly, each output has six adjusting
parameters. So the total number of adjusting parameters is to be thirty {5 (3 inputs + 2
outputs) × 2 (membership function) × 3 (adjusting parameters a , b , and c ) = 30}.
The ranges of adjusting parameters are defined as [ amin , amax ], [ bmin , bmax ] and [ cmin ,
cmax ] respectively, for lower and upper boundary of the WDO algorithm. The amin and
amax are 30 and 65 for the membership function of the inputs, respectively. The bmin and
bmax are 1 and 3.5 for the membership function of the inputs, respectively. The parameters
cmin and cmax are 20 and 150 for the membership function of inputs, respectively.
Similarly, the amin and amax are 2 and 5 for the membership function of outputs,
respectively. The bmin and bmax are 1 and 3.5 for the membership function of the outputs,
respectively. The parameters cmin and cmax are located at 6.7 and 16.7 for the membership
function of outputs, respectively. Figure 8.5 shows the air parcels representation of the
WDO algorithm. The optimized membership functions of the inputs ( d f , d l , and d r ) and
the outputs ( mr , and ml ) are shown in Figures 8.6 and 8.7, respectively. The results of
the adjusting parameters ( a , b , and c ) of the inputs and outputs after optimization are
listed in Table 8.4 and Table 8.5, respectively.
The most important step in applying the WDO algorithm is to select the fitness
function, which is used to evaluate the optimum pressure of the air parcels. In during the
optimization process, the combined root mean square errors (CRMSE) are used to
evaluate the fitness of the fuzzy controller: -
1 z
  mrActual  mrFW 
2
RMSE mr  (8.9)
z p 1
1 z
  mlActual  mlFW 
2
RMSE ml  (8.10)
z p 1
CRMSE=RMSEmr  RMSEml (8.11)
147
where mractual and mlactual are the actual value of right and left motor velocity, respectively.
The mrFW , and mlFW are calculated value of the right and left motor velocity, respectively
through Fuzzy-WDO. The RMSE mr , and RMSE ml are the root mean square error of the
right and left motor velocity respectively; and z is the iteration number.
a b c a b c a b c a b c
Input ( ) Input ( )
a b c a b c
Input ( )
a b c a b c a b c a b c
Output ( ) Output ( )
Figure 8.5: Air parcels representation of the WDO algorithm.
Table 8.4: Adjusting parameters of the inputs after optimization
Membership
Inputs a b c
function
Near 55.11 2.14 25

df
Far 59.6 1.88 149.4
Near 58.3 2.44 22.4

dl
Far 62.41 1.76 148.3
Near 57.42 2.33 23.1

dr
Far 60.29 1.55 148.9
148
Table 8.5: Adjusting parameters of the outputs after optimization
Membership
Outputs a b c
function
Low 3.61 2.601 6.515

mr
High 4.22 2.14 16.2
Low 3.97 2.21 5.96

ml
High 4.32 2.96 16.4
Figure 8.6: Fuzzy membership functions for the inputs ( d f , d l , and d r ) after
optimization.
149
Figure 8.7: Fuzzy membership functions for the outputs ( mr , and ml ) after optimization.
8.4 Simulation Results
This section describes the simulation results using S-Fuzzy and Fuzzy-WDO controllers
in the various unknown static and dynamic environments. The simulations are conducted
using the MATLAB software on the HP 3.40 GHz processor. Figures 8.8 and 8.9 show
the navigation result of the mobile robot between the obstacles and walls respectively,
using the S-Fuzzy and Fuzzy-WDO controller in the unknown environments. Similarly,
the Figure 8.10 demonstrates the navigation of a mobile robot in an unknown
environment with the presence of two dynamic obstacles using Fuzzy-WDO controller. It
is assumed that the position of the start point and goal point are known. But the positions
of all the obstacles in the environment are unknown for the robot. In the simulation
results, the green and red color trajectory indicates the path generated by the S-Fuzzy and
Fuzzy-WDO controllers, respectively. Simulation results show the Fuzzy-WDO
controller gives smooth and optimal path compared to the S-Fuzzy controller. Table 8.6
shows the navigation path length and time taken by the robot using the S-Fuzzy and
Fuzzy-WDO controller in the various unknown environments.
150
(a)
(cm)
(cm)
(b)
(cm)
(cm)
Figure 8.8: Mobile robot navigation between the obstacles using (a) S-Fuzzy and (b)
Fuzzy-WDO controller.
151
(a)
(cm)
(cm)
(b)
(cm)
(cm)
Figure 8.9: Mobile robot navigation between the walls using (a) S-Fuzzy and (b) Fuzzy-
WDO controller.
152
(i) (ii)
(cm)
(cm)
(cm) (cm)
(iii) (iv)
(cm)
(cm)
(cm) (cm)
Figure 8.10: Mobile robot navigation in the dynamic environment using Fuzzy-WDO
controller.
153
Table 8.6: The simulation results of S-Fuzzy and Fuzzy-WDO controllers
Navigation path Navigation time

Figure no. Controller
length (cm) (sec)
S-Fuzzy 79 7.2
Figure 8.8
Fuzzy-WDO 74 6.9
S-Fuzzy 104 9.1

Figure 8.9
Fuzzy-WDO 98 8.7
This section describes the computer simulation result comparison between the previous
model [45] and proposed Fuzzy-WDO controller in the same path planning problems. In
the article [45], the authors have used two simple fuzzy controllers such as tracking fuzzy
logic control (TFLC) and obstacle avoidance fuzzy logic control (OAFLC) without
adjusting its membership function for mobile robot navigation. Figures 8.11 and 8.12
show the mobile robot navigation in the same environment without obstacle using fuzzy
controller [45] and proposed Fuzzy-WDO controller, respectively. Similarly, Figures 8.13
and 8.14 present the path covered by the robot in the same environment with the four
obstacles using fuzzy controllers [45] and proposed Fuzzy-WDO controller, respectively.
From the simulation figures, it can be seen that the proposed Fuzzy-WDO controller
covers shorter distance to reach the goal as compared to previous model [45] because
WDO algorithm adjusts the membership function of the fuzzy controller, which provides
better result compared to the standalone fuzzy model. Table 8.7 illustrates the path traced
(in cm) by the robot to reach the goal using proposed controller and previous model [45].
The centimeter measurements are taken on the proportional basis.
154
Figure 8.11: Mobile robot navigation in an environment without obstacle using fuzzy
controller [45].
(cm)
(cm)
Figure 8.12: Mobile robot navigation in an environment without obstacle using Fuzzy-
WDO controller.
155
Figure 8.13: Mobile robot navigation in an environment with four obstacles using fuzzy
controller [45].
(cm)
(cm)
Figure 8.14: Mobile robot navigation in an environment with four obstacles using Fuzzy-
WDO controller.
156
Table 8.7: The simulation result comparison between the fuzzy controller [45] and
proposed Fuzzy-WDO controller

Figure no. Method
(cm)
Figure 8.11 Fuzzy controller [45] 181
Figure 8.12 Fuzzy-WDO controller 165
Figure 8.13 Fuzzy controller [45] 183
Figure 8.14 Fuzzy-WDO controller 173
8.6 Experimental Results
8.6.1 Khepera-III Mobile Robot Description
The experiments are conducted using the Khepera-III mobile robot in unknown
environments. The Khepera-III mobile robot has two wheels controlled by two DC servo
motors and one caster wheel. The diameter and height of the robot are 13cm and 7cm
respectively. The Khepera-III mobile robot is equipped with nine infrared proximity
sensors and five ultrasonic sensors, as shown in Figure 8.15. The infrared proximity
sensor reads obstacles up to 30cm, and the ultrasonic sensor reads obstacles from 20cm to
4m approximately. In this study, we have set the minimum and maximum velocity of
Khepera-III mobile robot between the 6.7-16.7cm/sec.
8.6.2 Experiments
In the experiments, the controllers are implemented in the Khepera-III mobile robot using
HP laptop. The width and height of the experimental platform are 250cm and 250cm,
respectively. Figures 8.16 and 8.17 show the real time navigation of the Khepera-III
mobile robot in unknown environments with the obstacles and walls, respectively. In
Figure 8.16, the start position of the robot is (175, 100) cm and the position of the goal is
(0, 200) cm. The starting angle between the robot and the goal is 29.74°. In Figure 8.17,
the start position of the robot is (50, 50) cm and the goal position is (250, 200) cm. The
starting angle between the robot and the goal is 36.87°. In the experiments, it is assumed
157
that the position of the start point and goal point are known, but the positions of all the
obstacles in the environment are unknown for the robot. The S-Fuzzy and Fuzzy-WDO
controller generates the motor velocity control commands of the robot for obstacle
avoidance using on-board sensor information. The successful experimental results in the
various unknown environments are shown below to verify the effectiveness of the S-
Fuzzy and Fuzzy-WDO controllers. Table 8.8 shows the experimental path length and
time taken by the Khepera-III mobile robot to reach target using the S-Fuzzy and Fuzzy-
WDO controllers in the various unknown environments. Tables 8.9 and 8.10 present the
experimental results, respectively. In the comparison study between the simulation and
experiments, it is observed that errors within 6% have been found, these are happened
due to slippage and friction during real time experiment.
Figure 8.15: Infrared proximity sensor distribution of Khepera-III mobile robot.
158
Goal
Start
(i) (ii)
Goal
Start
Y-AXIS
(iii) (iv)
X-AXIS
Figure 8.16: Real time navigation of Khepera-III mobile robot between the obstacles
using S-Fuzzy and Fuzzy-WDO controller.
159
Goal
Start
(i) (ii)
Goal
Y-AXIS
Start
(iii) (iv)
X-AXIS
Figure 8.17: Real time navigation of Khepera-III mobile robot between the walls using S-
Fuzzy and Fuzzy-WDO controller.
160
Table 8.8: The experimental results of S-Fuzzy and Fuzzy-WDO controllers
Experimental path Experimental

Figure no. Controller
length (cm) time (sec)
S-Fuzzy 84 7.7
Figure 8.16
Fuzzy-WDO 78 7.3
S-Fuzzy 111 9.7

Figure 8.17
Fuzzy-WDO 103 9.2
results
Figure no. Error between

Controller
experimental Simulation Experimental experimental
res.) result result results
Figure 8.8 and S-Fuzzy 79 84 5.95%

8.16 Fuzzy-WDO 74 78 5.12%
Figure 8.9 and S-Fuzzy 104 111 6.31%

8.17 Fuzzy-WDO 98 103 4.85%
161
Figure no. Error between

Controller
experimental Simulation Experimental experimental
res.) result result results
Figure 8.8 and S-Fuzzy 7.2 7.7 6.49%

8.16 Fuzzy-WDO 6.9 7.3 5.47%
Figure 8.9 and S-Fuzzy 9.1 9.7 6.18%

8.17 Fuzzy-WDO 8.7 9.2 5.43%
8.7 Summary
In this chapter, the two methods S-Fuzzy controller and the hybrid Fuzzy-WDO
algorithm have been applied to the mobile robot navigation. A new population-based
optimization algorithm, called Wind Driven Optimization (WDO) is used to optimize and
set the antecedent and consequent parameters of the fuzzy controller. The proposed
algorithms are successfully verified through simulations and real-time experiments in the
different environments. Simulation and experimental results demonstrate that the Fuzzy-
WDO controller provide better performance as compared to the S-Fuzzy controller. In the
comparison study between the simulation and experiment results errors are recorded, and
the errors are found due to the effect of slippage and friction between the wheels of the
robot and surface during navigation in real time mode. During experiment utmost care
has been taken to minimize the slippage and friction between the wheels and surface. Still
the effect of slippage and friction are unavoidable, and errors are recorded during the
comparison of the results for travelling path length (5.56%) and for navigation time
(5.9%).
162
Chapter 9
Comparative Study of the Proposed Soft

Computing Techniques Applied for
Mobile Robot Navigation
9.1 Introduction
In the previous chapters, the various soft computing techniques such as H-Fuzzy
architecture, CN-Fuzzy architecture, Fuzzy-SA algorithm, WDO algorithm, and Fuzzy-
WDO algorithm have been successfully designed and implemented for the mobile robot
navigation and obstacle avoidance in the different static and dynamic environments. This
chapter presents the comparative analysis of the all developed techniques in the two
different simulation and experimental test environments. In the comparative study, it is
found that the Fuzzy-WDO algorithm is more efficient (in terms of path length and
navigation time) as compared to rest of the techniques for mobile robot navigation. Rest
of the chapter is organized as follows: Section 9.2 presents the simulation result
comparison of the all developed soft computing techniques. Section 9.3 demonstrates the
experimental verification of the developed simulations. Finally, Section 9.4 depicts the
summary.
9.2 Simulation studies
All the simulations are performed through MATLAB software in the HP 3.40 GHz
processor personal computer. The simulation studies are divided into two test
environments: test1 and test2, respectively, which are presented below.
163
9.2.1 Simulation Test1
In the first simulation test, all the developed soft computing techniques are employed for
mobile robot navigation and obstacle avoidance in the environment. The dimensions of
the environment are 250cm width and 250cm height. Figures 9.1 (a) to 9.1 (e) show the
best results of the mobile robot navigation between the obstacles using H-Fuzzy
WDO algorithm, respectively. The start position of the mobile robot is (175, 100) cm and
the position of the goal is (0, 200) cm.
9.2.2 Simulation Test2
In the second simulation test, the mobile robot navigates between the walls in the
environment using all developed soft computing techniques. The width and height of the
environment are 270cm and 240cm, respectively. Figures 9.2 (a) to 9.2 (e) illustrate the
best results of the mobile robot navigation in the test environment using H-Fuzzy
WDO algorithm, respectively. The start and goal position of the mobile robot are (50, 50)
cm and (250, 200) cm, respectively.
In both the simulation tests (test1 and test2), it is assumed that the position of the start
point and goal point are known, but the positions of the obstacles and walls are unknown
for the robot. A minimum threshold distance has been fixed between the robot and the
obstacles. If the robot detects the obstacles in the threshold range, then the proposed
techniques provide the motion and orientation control command to the robot for avoiding
the obstacle and wall.
From both the simulation tests, it is observed that the Fuzzy-WDO algorithm covers
shorter distance to reach the goal as compared to the rest of the techniques because WDO
algorithm adjusts the membership function of the fuzzy controller, which provides better
result compared to the other techniques. For each simulation test, we have run all the
developed techniques for 10-15 times, and the average path length and time have been
listed in the table. Table 9.1 shows the average navigation path length (in cm) and time
taken (in the sec) by the robot to reach the goal using all the developed techniques in
different environments.
164
(cm)
(cm)
(cm) (cm)
Figure 9.1 (a): Figure 9.1 (b):
(cm)
(cm)
(cm) (cm)
Figure 9.1 (c): Figure 9.1 (d):
Figure 9.1 (a): H-Fuzzy architecture.

Figure 9.1 (b): CN-Fuzzy architecture.
(cm)
Figure 9.1 (c): Fuzzy-SA algorithm.

Figure 9.1 (d): WDO algorithm.
Figure 9.1 (e): Fuzzy-WDO algorithm.
(cm)
Figure 9.1 (e):
Figure 9.1: Mobile robot navigation and obstacle avoidance in the simulation test1 using
the developed soft computing techniques.
165
(cm)
(cm)
(cm) (cm)
Figure 9.2 (a): Figure 9.2 (b):
(cm)
(cm)
(cm) (cm)
Figure 9.2 (c): Figure 9.2 (d):

(cm)

(cm)
Figure 9.2 (e):
Figure 9.2: Mobile robot navigation and obstacle avoidance in the simulation test2 using
the developed soft computing techniques.
166
Table 9.1: Simulation results of the mobile robot navigation in the test1 and test2 using
all developed techniques
Navigation path Navigation time

Figure no. Techniques
length (cm) (sec)
H-Fuzzy 80 7.7
CN-Fuzzy 78 7.5
Figure 9.1 (a) to
(e) Fuzzy-SA algorithm 76 7.4
(Test1)
WDO algorithm 75 7.2
Fuzzy-WDO algorithm 74 6.9
H-Fuzzy 105 9.4
CN-Fuzzy 102 9.3

Figure 9.2 (a) to
(e) Fuzzy-SA algorithm 101 9.2
(Test2)
WDO algorithm 100 8.9
Fuzzy-WDO algorithm 98 8.7
Note: Bold value indicates the best results compared to rest of the techniques.
9.3 Experimental studies of the developed simulations
This section demonstrates the experimental verification of the developed simulations in

test1 and test2. All the experiments have been conducted through C/C++ running Arduino
microcontroller based wheeled mobile robot. The other descriptions of the experimental
mobile robot have been discussed in the previous chapters and appendix. The size of the
experimental platform is 300cm width and 300cm height. The experimental studies are
also divided into two test environments: test1 and test 2, respectively, which is presented
below.
167
9.3.1 Experimental Test1
In the experimental test1, all the developed soft computing techniques have experimented
between the four square shaped obstacles. The robot moves from the start point (175cm,
100cm) to goal point (0cm, 200cm). The starting angle between the robot and the goal is
29.74°. Figures 9.3 (a) to 9.3 (e) present the experimental verification of the above
simulation results (see the Figures 9.1) using H-Fuzzy architecture, CN-Fuzzy
architecture, Fuzzy-SA algorithm, WDO algorithm, and Fuzzy-WDO algorithm,
respectively. From the experimental results, it can be observed that the robot successfully
avoids all the obstacles and reach the goal similar to simulation test1.
9.3.2 Experimental Test2
In the experimental test2, the mobile robot navigates from the start point (50cm, 50cm) to
goal point (250cm, 200cm) between the walls in the environment using all developed soft
computing techniques. The starting angle between the robot and the goal is 36.87°.
Figures 9.4 (a) to 9.4 (e) show the real-time navigation results of the wheeled mobile
robot using H-Fuzzy architecture, CN-Fuzzy architecture, Fuzzy-SA algorithm, WDO
algorithm, and Fuzzy-WDO algorithm, respectively. In the experimental test2, it can be
observed that the trajectory of mobile robot navigation is similar to the trajectory of
simulation test2.
In the experimental studies (test1 and test2), we have set a threshold distance (30cm)
between the robot and the obstacles. If the robot detects the obstacles in the threshold
range, then the developed techniques send the motion and orientation control command to
the robot for avoiding the obstacle.
For each experimental test, we have run all the developed techniques for 10-15 times,
and the average path length and time have been listed in the table. Table 9.2 illustrates the
real-time average navigation path length (in cm) and time taken (in the sec) by the robot
to reach the goal using developed techniques in the different environments. From the
experimental studies, it can be concluded that the Fuzzy-WDO algorithm covers shorter
distance to reach the goal as compared to other developed techniques.
168
Goal Goal
Start Start
Y-AXIS
Figure 9.3 (a): Y-AXIS Figure 9.3 (b):

X-AXIS X-AXIS
Goal Goal
Start Start
Y-AXIS
Y-AXIS
X-AXIS Figure 9.3 (c): X-AXIS Figure 9.3 (d):
169
Goal
Start Figure 9.3 (d): WDO algorithm.
Y-AXIS
X-AXIS Figure 9.3 (e):
Figure 9.3: Mobile robot navigation and obstacle avoidance in the experimental test1
using the developed soft computing techniques.
Goal Goal
Y-AXIS
Y-AXIS
Start Start
X-AXIS Figure 9.4 (a): X-AXIS Figure 9.4 (b):
170
Goal Goal
Y-AXIS
Y-AXIS
Start Start
X-AXIS Figure 9.4 (c): X-AXIS Figure 9.4 (d):
Goal

Y-AXIS
Start
X-AXIS Figure 9.4 (e):
Figure 9.4: Mobile robot navigation and obstacle avoidance in the experimental test2
using the developed soft computing techniques.
171
Table 9.2: Experimental results of the mobile robot navigation in the test1 and test2 using
all developed techniques

Figure Navigation time (sec)
Techniques (cm)
no.
Simulation Experimental Simulation Experimental
H-Fuzzy 80 85 7.7 8.2
CN-Fuzzy 78 83 7.5 8.0
Figure 9.3 Fuzzy-SA

76 81 7.4 7.8
(a) to (e) algorithm
(Test1)
WDO algorithm 75 80 7.2 7.6
Fuzzy-WDO
74 78 6.9 7.2
algorithm
H-Fuzzy 105 112 9.4 10
CN-Fuzzy 102 109 9.3 9.9
Figure 9.4 Fuzzy-SA

101 107 9.2 9.8
(a) to (e) algorithm
(Test2)
WDO algorithm 100 105 8.9 9.4
Fuzzy-WDO
98 103 8.7 9.1
algorithm
Note: Bold value indicates the best results compared to rest of the techniques.
172
9.4 Summary
In the present study, all the developed soft computing techniques such as H-Fuzzy
WDO algorithm are successfully employed for mobile robot navigation and obstacle
avoidance in the different simulation and experimental environments. Firstly, the two
simulation tests are conducted using all the developed techniques to compare their
performances in the test environments. Then, the experiments have been done in the same
environment to verify the effectiveness of the developed simulations. In the comparison
study between the simulation and experiment results errors are recorded, and the errors
are found due to the effect of slippage and friction between the wheels of the robot and
surface during navigation in real time mode.
During experiment utmost care has been taken to minimize the slippage and friction
between the wheels and surface. Still the effect of slippage and friction are unavoidable,
and errors are recorded during the comparison of the results for travelling path length
(6%) and for navigation time (7%).
After summarizing the above simulation and experimental studies, it can be
concluded that the Fuzzy-Wind Driven Optimization algorithm gives the better results (in
terms of path length and navigation time) as compared to rest of the techniques, which
verifies the superiority of this technique.
173
Chapter 10
Conclusion and Scope for Future

Research
10.1 Introduction
In this proposed research work, various soft computing techniques such as Hybrid Fuzzy
(H-Fuzzy) architecture, Cascade Neuro-Fuzzy (CN-Fuzzy) architecture, Fuzzy-Simulated
Annealing (Fuzzy-SA) algorithm, Wind Driven Optimization (WDO) algorithm, and
Fuzzy-Wind Driven Optimization (Fuzzy-WDO) algorithm have been successfully
designed and implemented for mobile robot navigation and obstacle avoidance in the
different static and dynamic environments. This chapter summarizes the significant
contributions, conclusion, and scope for the future research from proposed research work.
Rest of the chapter is organized as follows: Section 10.2 presents the important
contributions of the proposed dissertation. Section 10.3 demonstrates the conclusions of
all chapters. Section 10.4 depicts the scope for future research from the present research.
10.2 Important contributions
The significant contributions of the dissertation are summarized as follows: -
 The kinematic and dynamic equations of the mobile robot have been described in
the third chapter, which helps to control the motion and orientation of the robot in
the environment.
 The Hybrid Fuzzy (H-Fuzzy) architecture, Cascade Neuro-Fuzzy (CN-Fuzzy)
architecture, Fuzzy-Simulated Annealing (Fuzzy-SA) algorithm have been applied
in this research work to improve the navigation and obstacle avoidance
performance of the mobile robot in various (static and dynamic) environments.
174
 According to literature survey, this is for the first time that Wind Driven
Optimization (WDO) algorithm has been applied for the mobile robot navigation
and obstacle avoidance in the static and dynamic environments.
 Besides, this newly developed WDO algorithm is integrated with the fuzzy
controller to adjust and optimize the antecedent and consequent parameters of the
fuzzy membership function.
 The performance of these developed techniques are demonstrated through
computer simulations using MATLAB software and implemented in real time by
using Arduino microcontroller based experimental mobile robots.
10.3 Conclusions
According to the results obtained from the simulation and experimental studies using all
the developed soft computing techniques, the major conclusions of the dissertation are
summarized below: -
 The kinematic and dynamic analysis of the mobile robot have been presented in
the third chapter, which is important for controlling the right and left wheel
velocities of the mobile robot during navigation and obstacle avoidance in the
environment.
 Chapter four discusses the design and implementation of the hybrid fuzzy (H-
Fuzzy) architecture for mobile robot navigation in the presence of static and
dynamic obstacles. The proposed H-Fuzzy architecture aims to control the turning
angle (between the robot and goal) and motor velocities of a mobile robot from
starting point to goal point with obstacle avoidance competence. This architecture
is tested in various simulation and experimental environments and is found to be a
good agreement.
 In chapter five, the Cascade Neuro-Fuzzy (CN-Fuzzy) architecture has been
presented for intelligent navigation control of a mobile robot in the static and
dynamic environments filled with obstacles. The cascade neural network is used
to train the robot to reach the goal in the environment, and the fuzzy logic
architecture is integrated with this cascade neural network to control the right and
left motor velocities of the robot to protect it from the obstacles.
175
 In chapter six, a Takagi-Sugeno fuzzy model is integrated with the Simulated
Annealing algorithm called as Fuzzy-Simulated Annealing (Fuzzy-SA) algorithm
to optimize the navigation path length of the mobile robot in the given
environment. The proposed algorithm receives the obstacle distance (inputs) from
the group of sensors for selecting the suitable steering angle (output) control
command for the mobile robot during navigation. Successful simulation and
experimental studies verify the effectiveness and efficiency of this proposed
Fuzzy-SA algorithm.
 Chapter seven and eight applies the Wind Driven Optimization (WDO) algorithm,
and Fuzzy-Wind Driven Optimization (Fuzzy-WDO) to solve the optimal path
planning problems of a mobile robot in different simulation and experimental
environments. The WDO (Wind Driven Optimization) algorithm has been used to
optimize and tune the input/output membership function parameters of the fuzzy
controller. These algorithms improve the navigation performance of the mobile
robot in the given environments and produce a smooth navigation path within a
reasonable time.
 In chapter nine, the two simulation and experimental tests (test1 and test2) are
conducted to compare the performance of all the developed soft computing
techniques in the given environments. From the comparison tests, it is found that
as compared to the H-Fuzzy, CN-Fuzzy, Fuzzy-SA, and WDO the Fuzzy-WDO
provides better results (in terms of path length and navigation time) because WDO
algorithm adjusts the membership function of the fuzzy controller, which verifies
the superiority of this newly developed technique.
 During comparison studies between simulation and experiment, the average
percentage of errors are found to be within 6% in terms of travelling path length
and 7% in terms of navigation time.
 All the simulation results, experimental studies, and comparison with previous
works (e.g., [11] and [45]) illustrate that the proposed soft computing techniques
are indeed effective and feasible for navigation of mobile robot form start to goal
in a cluttered environment.
176
10.4 Scope for Future Research
The present research work can be extended in a number of ways and some of these are
listed below: -
 The proposed techniques are developed and tested for the navigation of single
robot in dynamic environments. In future, these developed techniques will be
extended for multi-robot navigation and obstacle avoidance in the dynamic
environments.
 These developed soft computing techniques can be hybridized with other nature-
inspired optimization algorithms such as Teaching-Learning-Based Optimization
(TLBO), Firefly Algorithm (FA), Cuckoo Search (CS) algorithm, and Bat
Algorithm, etc. to compare the navigational performances and obstacle avoidance
strategy of the mobile robot.
 The proposed techniques may be implemented for the multiple swarms mobile
robots navigation and obstacle avoidance with moving goal problems in the
environment.
 The vision sensors like camera may be attached to the experimental mobile robot
to improve the sensory system of the mobile robot for the betterment of the
navigation accuracy.
 In future research work, these developed techniques will be tested in the rough
terrain to study its performance.
177
Bibliography
[1] Ni J., Wu L., Fan X. and Yang S. X. Bioinspired Intelligent Algorithm and Its Applications
for Mobile Robot Control: A Survey. Computational Intelligence and Neuroscience, Article
ID 3810903: 1--15, 2016.
[2] Takahashi O. and Schilling R.J. Motion Planning in a Plane Using Generalized Voronoi
Diagrams. IEEE Transactions on Robotics and Automation, 5(2): 143--150, 1989.
[3] Bhattacharya P. and Gavrilova M. L. Roadmap-Based Path Planning-Using the Voronoi

Diagram for a Clearance-Based Shortest Path. IEEE Robotics and Automation Magazine,
15(2): 58--66, 2008.
[4] Gomez E. J., Martinez Santa F. and Martinez Sarmiento F. H. A Comparative Study of
Geometric Path Planning Methods for a Mobile Robot: Potential Field and Voronoi
Diagrams. In IEEE International Congress of Engineering Mechatronics and Automation
(CIIMA), Bogota, Colombia, pages 1--6, 2013.
[5] Abiyev R., Ibrahim D. and Erin B. Navigation of Mobile Robots in the Presence of
Obstacles. ELSEVIER Advances in Engineering Software, 41(10): 1179--1186, 2010.
[6] Soltani A. R., Tawfik H., Goulermas J. Y. and Fernando T. Path Planning in Construction
Sites: Performance Evaluation of the Dijkstra, A∗, and GA Search Algorithms. ELSEVIER
Advanced Engineering Informatics, 16(4): 291--303, 2002.
[7] Masehian E. and Amin-Naseri M. R. A Voronoi Diagram–Visibility Graph–Potential Field

Compound Algorithm for Robot Path Planning. Journal of Robotic System, 21(6): 275--
300, 2004.
[8] Weigl M., Siemiaatkkowska B., Sikorski K. A. and Borkowski A. Grid-Based Mapping for
Autonomous Mobile Robot. ELSEVIER Robotics and Autonomous Systems, 11(1): 13--21,
1993.
[9] Park K. H., Kim Y. J. and Kim J. H. Modular Q-Learning based Multi-Agent Cooperation
for Robot Soccer. ELSEVIER Robotics and Autonomous Systems, 35(2): 109--122, 2001.
178
[10] Montaner M. B. and Ramirez-Serrano A. Fuzzy Knowledge-Based Controller Design for
Autonomous Robot Navigation. ELSEVIER Expert Systems with Applications, 14(1): 179--
186, 1998.
[11] Engedy I. and Horvath G. Artificial Neural Network based Local Motion Planning of a
Wheeled Mobile Robot. In IEEE International Symposium on Computational Intelligence
and Informatics (CINTI), pages 213--218, 2010.
[12] Zhu A. and Yang S. X. Neurofuzzy-Based Approach to Mobile Robot Navigation in

Unknown Environments. IEEE Transactions on Systems, Man, and Cybernetics
Applications and Reviews, 37(4): 610--621, 2007.
[13] Ghorbani A., Shiry S. and Nodehi A. Using Genetic Algorithm for a Mobile Robot Path
Planning. In IEEE International Conference on Future Computer and Communication,
Kuala Lumpur, Malaysia, pages 164--166, 2009.
[14] Ahmadzadeh S. and Ghanavati M. Navigation of Mobile Robot Using the PSO Particle
Swarm Optimization. Journal of Academic and Applied Studies (JAAS), 2(1): 32--38, 2012.
[15] Garcia M. P., Montiel O., Castillo O., Sepulveda R. and Melin P. Path Planning for
Autonomous Mobile Robot Navigation with Ant Colony Optimization and Fuzzy Cost
Function Evaluation. ELSEVIER Applied Soft Computing, 9(3): 1102--1110, 2009.
[16] Miao H. and Tian Y. C. Dynamic Robot Path Planning using an Enhanced Simulated
Annealing Approach. ELSEVIER Applied Mathematics and Computation, 222: 420--437,
2013.
[17] Hui N. B., Mahendar V. and Pratihar D. K. Time-Optimal, Collision-Free Navigation of a

Car-Like Mobile Robot using Neuro-Fuzzy Approaches. ELSEVIER Fuzzy Sets and
Systems, 157(16): 2171--2204, 2006.
[18] Abadi D. N. M. and Khooban M. H. Design of Optimal Mamdani-Type Fuzzy Controller

for Nonholonomic Wheeled Mobile Robots. ELSEVIER Journal of King Saud University-
Engineering Sciences, 27(1): 92--100, 2015.
[19] Chakraborty N. and Ghosal A. Kinematics of Wheeled Mobile Robots on Uneven Terrain.
ELSEVIER Mechanism and Machine Theory, 39(12): 1273--1287, 2004.
179
[20] Wang X. and Yang S. X. A Neuro-Fuzzy Approach to Obstacle Avoidance of a
Nonholonomic Mobile Robot. In IEEE/ASME International Conference on Advanced
Intelligent Mechatronics, pages 29--34, 2003.
[21] Das T., Kar I. N. and Chaudhury S. Simple Neuron-Based Adaptive Controller for a
Nonholonomic Mobile Robot Including Actuator Dynamics. ELSEVIER Neurocomputing,
69(16): 2140--2151, 2006.
[22] Martinez R., Castillo O. and Aguilar L. T. Optimization of Interval Type-2 Fuzzy Logic
Controllers for a Perturbed Autonomous Wheeled Mobile Robot using Genetic Algorithms.
ELSEVIER Information Sciences, 179(13): 2158--2174, 2009.
[23] Al-Araji A. S., Abbod M. F. and Al-Raweshidy H. S. Applying Posture Identifier in

Designing an Adaptive Nonlinear Predictive Controller for Nonholonomic Mobile Robot.
ELSEVIER Neurocomputing, 99: 543--554, 2013.
[24] Liang Y., Xu L., Wei R. and Hu H. Adaptive Fuzzy Control for Trajectory Tracking of
Mobile Robot. In IEEE/RSJ International Conference on Intelligent Robots and Systems
(IROS), pages 4755--4760, 2010.
[25] Zadeh L. A. The Concept of a Linguistic Variable and its Application to Approximate
Reasoning-I. ELSEVIER Information Sciences, 8(3): 199--249, 1975.
[26] Ren L., Wang W. and Du Z. A New Fuzzy Intelligent Obstacle Avoidance Control Strategy
for Wheeled Mobile Robot. In IEEE International Conference on Mechatronics and
Automation (ICMA), pages 1732--1737, 2012.
[27] Yousfi N., Rekik C., Jallouli M. and Derbel N. Optimized Fuzzy Controller for Mobile
Robot Navigation in a Cluttered Environment. In IEEE 7th International Multi-Conference
on Systems, Signals and Devices, pages 1--7, 2010.
[28] Qing-yong B., Shun-ming L., Wei-yan S. and Mu-jin A. A Fuzzy Behavior-Based
Architecture for Mobile Robot Navigation in Unknown Environments. In IEEE
International Conference on Artificial Intelligence and Computational Intelligence, pages
257--261, 2009.
[29] Boubertakh H., Tadjine M., Glorennec P. and Labiod S. A Simple Goal Seeking Navigation
Method for a Mobile Robot using Human Sense, Fuzzy Logic and Reinforcement Learning.
Journal of Automatic Control, University of Belgrade, 18(1): 23--27, 2008.
180
[30] Muthu T., Thierry Gloude R., Swaminathan S. and Satish Kumar L. Fuzzy Logic Controller
for Autonomous Navigation. In IEEE International Conference on Communications and
Signal Processing (ICCSP), pages 81--92, 2012.
[31] El-Teleity S. A. L., Nossair Z. B., Mansour H. M. and TagElDein A. Fuzzy Logic Control
of an Autonomous Mobile Robot. In IEEE 16th International Conference on Methods and
Models in Automation and Robotics (MMAR), pages 188--193, 2011.
[32] Raguraman S. M., Tamilselvi D. and Shivakumar N. Mobile Robot Navigation Using
Fuzzy logic Controller. In IEEE 16th International Conference on Control, Automation,
Communication and Energy Conservation, pages 1--5, 2009.
[33] Wu S., Li Q., Zhu E., Xie J. and Zhichao G. Fuzzy Controller of Pipeline Robot Navigation
Optimized by Genetic Algorithm. In IEEE International Conference on Control and
Decision Conference (CCDC), China, pages 904--908, 2008.
[34] Samsudin K., Ahmad F. A. and Mashohor S. A Highly Interpretable Fuzzy Rule Base using
Ordinal Structure for Obstacle Avoidance of Mobile Robot. ELSEVIER Applied Soft
Computing, 11(2): 1631--1637, 2011.
[35] Beom H. R. and Cho K. S. A Sensor-Based Navigation for a Mobile Robot using Fuzzy
Logic and Reinforcement Learning. IEEE Transactions on Systems, Man, and Cybernetics,
25(3): 464--477, 1995.
[36] Pradhan S. K., Parhi D. R. and Panda A. K. Fuzzy Logic Techniques for Navigation of
Several Mobile Robots. ELSEVIER Applied Soft Computing, 9(1): 290--304, 2009.
[37] Liu Q., Lu Y. G. and Xie C. X. Optimal Genetic Fuzzy Obstacle Avoidance Controller of
Autonomous Mobile Robot Based on Ultrasonic Sensors. In IEEE International Conference
on Robotics and Biomimetics, China, pages 125--129, 2006.
[38] Farooq U., Hasan K. M., Abbas G. and Asad M. U. Comparative Analysis of Zero Order
Sugeno and Mamdani Fuzzy Logic Controllers for Obstacle Avoidance Behavior in Mobile
Robot Navigation. In IEEE International Workshop on Current Trends in Information
Technology (CTIT), pages 113--119, 2011.
[39] Algabri M., Mathkour H., Ramdane H. and Alsulaiman M. Comparative Study of Soft
Computing Techniques for Mobile Robot Navigation in an Unknown Environment.
ELSEVIER Computers in Human Behavior, 50: 42--56, 2015.
181
[40] Hui N. B. and Pratihar D. K. A Comparative Study on Some Navigation Schemes of a Real
Robot Tackling Moving Obstacles. ELSEVIER Robotics and Computer-Integrated
Manufacturing, 25(4): 810--828, 2009.
[41] Rusu C. G., Birou I. T. and Szoke E. Fuzzy Based Obstacle Avoidance System for
Autonomous Mobile Robot. In IEEE International Conference on Automation Quality and
Testing Robotics (AQTR), pages 1--6, 2010.
[42] Abdessemed F., Benmahammed K. and Monacelli E. A Fuzzy-Based Reactive Controller

for a Non-Holonomic Mobile Robot. ELSEVIER Robotics and Autonomous Systems, 47(1):
31--46, 2004.
[43] Selekwa M. F., Dunlap D. D., Shi D. and Collins E. G. Robot Navigation in Very Cluttered
Environments by Preference-Based Fuzzy Behaviors. ELSEVIER Robotics and Autonomous
Systems, 56(3): 231--246, 2008.
[44] Pratihar D. K., Deb K. and Ghosh A. A Genetic-Fuzzy Approach for Mobile Robot
Navigation among Moving Obstacles. ELSEVIER International Journal of Approximate
Reasoning, 20(2): 145--172, 1999.
[45] Faisal M., Hedjar R., Al Sulaiman M. and Al-Mutib K. Fuzzy Logic Navigation and
Obstacle Avoidance by a Mobile Robot in an Unknown Dynamic Environment. INTECH
International Journal of Advanced Robotic Systems, 10(37): 1--7, 2013.
[46] Babalou A. and Seifiour N. Application of Fuzzy Decision Making in Mobile Robot
Navigation in Dynamic Environments. In IEEE International Conference on Fuzzy
Systems, Korea, pages 877--881, 2009.
[47] Li T. H. S., Chang S. J. and Chen Y. X. Implementation of Human-Like Driving Skills by

Autonomous Fuzzy Behavior Control on an FPGA-Based Car-Like Mobile Robot. IEEE
Transactions on Industrial Electronics, 50(5): 867--880, 2003.
[48] Li T. H. S., Chang S. J. and Tong W. Fuzzy Target Tracking Control of Autonomous
Mobile Robots by Using Infrared Sensors. IEEE Transactions on Fuzzy Systems, 12(4):
491--501, 2004.
[49] Dongshu W., Yusheng Z. and Wenjie S. Behavior-Based Hierarchical Fuzzy Control for
Mobile Robot Navigation in Dynamic Environment. In IEEE International Conference on
Control and Decision Conference (CCDC), pages 2419--2424, 2011.
182
[50] Antonelli G., Chiaverini S. and Fusco G. A Fuzzy-Logic-Based Approach for Mobile
Robot Path Tracking. IEEE Transactions on Fuzzy Systems, 15(2): 211--221, 2007.
[51] Ayari E., Hadouaj S. and Ghedira K. A Fuzzy Logic Method for Autonomous Robot
Navigation in Dynamic and Uncertain Environment Composed with Complex Traps. In
IEEE 5th International Multi-conference on Computing in the Global Information
Technology, pages 18--23, 2010.
[52] Zou A. M., Hou Z. G., Fu S. Y. and Tan M. Neural Networks for Mobile Robot Navigation:
a Survey. In: Advances in Neural Networks-ISNN, pages 1218--1226, 2006.
[53] Xiao H., Liao L. and Zhou F. Mobile Robot Path Planning Based on Q-ANN. In IEEE
International Conference on Automation and Logistics, China, pages 2650--2654, 2007.
[54] Rai N., and Rai B. Neural Network based Closed loop Speed Control of DC Motor using
Arduino Uno. International Journal of Engineering Trends and Technology, 4(2): 137--
140, 2013.
[55] Patino H. D. and Carelli R. Neural Network-Based Optimal Control for Autonomous
Mobile Vehicle Navigation. In IEEE International Symposium on Intelligent Control, pages
391--396, 2004.
[56] Yang S. X. and Meng M. Neural Network Approaches to Dynamic Collision-Free

Trajectory Generation. IEEE Transactions on Systems, Man, and Cybernetics Applications
and Reviews, 31(3): 302--318, 2001.
[57] Nichols E., McDaid L. J. and Siddique N. Biologically Inspired SNN for Robot Control.
IEEE Transactions on Cybernetics, 15(2): 115--128, 2013.
[58] Motlagh O., Nakhaeinia D., Tang S. H., Karasfi B. and Khaksar W. Automatic Navigation
of Mobile Robots in Unknown Environments. SPRINGER Neural Computing and
Applications, 24(7-8): 1569--1581, 2014.
[59] Gavrilov A. V. and Lee S. An Architecture of Hybrid Neural Network Based Navigation
System for Mobile Robot. In IEEE Seventh International Conference on Intelligent Systems
Design and Applications, pages 587--590, 2007.
183
[60] Singh M. K. and Parhi D. R. Path Optimisation of a Mobile Robot Using an Artificial
Neural Network Controller. International Journal of Systems Science, 42(1): 107--120,
2011.
[61] Rossomando F. G. and Soria C. M. Design and Implementation of Adaptive Neural PID for
Non Linear Dynamics in Mobile Robots. IEEE Latin America Transactions, 13(4): 913--
918, 2015.
[62] Al-Jarrah R., Shahzad A. and Roth H. Path Planning and Motion Coordination for Multi-
Robots System Using Probabilistic Neuro-Fuzzy. ELSEVIER IFAC-Papers Online, 48(10):
46--51, 2015.
[63] Janglova D. Neural Networks in Mobile Robot Motion. International Journal of Advanced
Robotic Systems, 1(1): 15--22, 2004.
[64] Glasius R., Komoda A. and Gielen S. C. Neural Network Dynamics for Path Planning and
Obstacle Avoidance. Neural Networks, 8(1): 125--133, 1995.
[65] Kim C. J. and Chwa D. Obstacle Avoidance Method for Wheeled Mobile Robots Using
Interval Type-2 Fuzzy Neural Network. IEEE Transactions on Fuzzy Systems, 23(3): 677--
687, 2015.
[66] Mahmud F., Arafat A. and Zuhori S. T. Intelligent Autonomous Vehicle Navigated by
Using Artificial Neural Network. In IEEE International Conference on Electrical and
Computer Engineering, Dhaka, Bangladesh, pages 105--108, 2012.
[67] Chohra A., Farah A. and Benmehrez C. Neural Navigation Approach for Intelligent
Autonomous Vehicles (IAV) in Partially Structured Environments. SPRINGER Applied
Intelligence, 8(3): 219--233, 1998.
[68] Brahmi H., Ammar B. and Alimi A. M. Intelligent Path Planning Algorithm For
Autonomous Robot Based on Recurrent Neural Networks. In IEEE International
Conference on Advanced Logistics and Transport (ICALT), pages 199--204, 2013.
[69] Yang S. X., Hu T., Yuan X., Liu P. X. and Meng M. A Neural Network based Torque
Controller for Collision-Free Navigation of Mobile Robots. In IEEE International
Conference on Robotics and Automation (ICRA), Taipei, Taiwan, pages 13--18, 2003.
184
[70] Al Mutib K. and Mattar E. Neuro-fuzzy Controlled Autonomous Mobile Robotics System.
In IEEE 13th International Conference on Modelling and Simulation, UK, pages 1--7,
2011.
[71] Godjevac J. and Steele N. Neuro-Fuzzy Control of a Mobile Robot. ELSEVIER

Neurocomputing, 28(1): 127--143, 1999.
[72] Li W., Ma C. and Wahl F. M. A Neuro-Fuzzy System Architecture for Behavior-Based

Control of a Mobile Robot in Unknown Environments. ELSEVIER Fuzzy Sets and Systems,
87(2): 133--140, 1997.
[73] Joshi M. M. and Zaveri M. Neuro-Fuzzy Based Autonomous Mobile Robot Navigation
System. In IEEE 11th International Conference Control, Automation, Robotics and Vision,
Singapore, pages 384--389, 2010.
[74] Marichal G. N., Acosta L., Moreno L., Mendez J. A., Rodrigo J. J. and Sigut M. Obstacle
Avoidance for a Mobile Robot: A Neuro-Fuzzy Approach. ELSEVIER Fuzzy Sets and
Systems, 124(2): 171--179, 2001.
[75] Zhang N., Beetner D., Wunsch D. C., Hemmelman B. and Hasan A. An Embedded Real-
Time Neuro-Fuzzy Controller for Mobile Robot Navigation. In IEEE International
Conference on Fuzzy Systems, pages 319--324, 2005.
[76] Baturone I., Gersnoviez A. and Barriga A. Neuro-Fuzzy Techniques to Optimize an FPGA
Embedded Controller for Robot Navigation. ELSEVIER Applied Soft Computing, 21: 95--
106, 2014.
[77] Ma X., Li X. and Qiao H. Fuzzy Neural Network-Based Real-Time Self-Reaction of

Mobile Robot in Unknown Environments. ELSEVIER Mechatronics, 11(8): 1039--1052,
2001.
[78] Imen M., Mohammad M. and Shoorehdeli M. A. Tracking Control of Mobile Robot Using
ANFIS. In IEEE International Conference on Mechatronics and Automation, Beijing,
China, pages 422--427, 2011.
[79] Ganapathy V., Yun S. C. and Ng J. Fuzzy and Neural Controllers for Acute Obstacle
Avoidance in Mobile Robot Navigation. In IEEE/ASME International Conference on
Advanced Intelligent Mechatronics, Singapore, pages 1236--1241, 2009.
185
[80] Zhao T. and Wang Y. A Neural-Network based Autonomous Navigation System using
Mobile Robots. In IEEE 12th International Conference on Control, Automation, Robotics
& Vision (ICARCV), Guangzhou, China, pages 1101--1106, 2012.
[81] Kumar D. and Dhama K. Neuro-Fuzzy Control of an Intelligent Mobile Robot. In: IEEE
Second International Conference on Advanced Computing & Communication Technologies
(ACCT), pages 106--111, 2012.
[82] Song K. T. and Sheen L. H. Heuristic Fuzzy-Neuro Network and its Application to
Reactive Navigation of a Mobile Robot. ELSEVIER Fuzzy Sets and Systems, 110(3): 331--
340, 2000.
[83] Lee C. H. and Chiu M. H. Recurrent Neuro Fuzzy Control Design for Tracking of Mobile
Robots via Hybrid Algorithm. ELSEVIER Expert Systems with Applications, 36(5): 8993--
8999, 2009.
[84] Deshpande S. U. and Bhosale S. S. Adaptive Neuro-Fuzzy Inference System based Robotic
Navigation. In IEEE International Conference on Computational Intelligence and
Computing Research (ICCIC), pages 1--4, 2013.
[85] Rusu P., Petriu E. M., Whalen T. E., Cornell A. and Spoelder H. J. Behavior-Based Neuro-
Fuzzy Controller for Mobile Robot Navigation. IEEE Transactions on Instrumentation and
Measurement, 52(4): 1335--1340, 2003.
[86] Pothal J. K. and Parhi D. R. Navigation of Multiple Mobile Robots in a Highly Clutter
Terrains using Adaptive Neuro-Fuzzy Inference System. ELSEVIER Robotics and
Autonomous Systems, 72: 48--58, 2015.
[87] Ng K. C. and Trivedi M. M. A Neuro-Fuzzy Controller for Mobile Robot Navigation and
Multirobot Convoying. IEEE Transactions on Systems, Man, and Cybernetics, 28(6): 829--
840, 1998.
[88] Demirli K. and Khoshnejad M. Autonomous Parallel Parking of a Car-Like Mobile Robot
by a Neuro-Fuzzy Sensor-Based Controller. ELSEVIER Fuzzy Sets and Systems, 160(19):
2876--2891, 2009.
[89] Al-Mayyahi A., Wang W. and Birch P. Adaptive Neuro-Fuzzy Technique for Autonomous
Ground Vehicle Navigation. Robotics, 3(4): 349--370, 2014.
186
[90] Pradhan S. K., Parhi, D. R. and Panda A. K. Neuro-Fuzzy Technique for Navigation of
Multiple Mobile Robots. SPRINGER Fuzzy Optimization and Decision Making, 5(3): 255--
288, 2006.
[91] Algabri M., Mathkour H. and Ramdane H. Mobile Robot Navigation and Obstacle-
Avoidance Using ANFIS in Unknown Environment. International Journal of Computer
Applications, 91(14): 36--41, 2014.
[92] Elshamli A., Abdullah H. A. and Areibi S. Genetic Algorithm for Dynamic Path Planning.
In IEEE International Conference on Electrical and Computer Engineering, pages 677--
680, 2004.
[93] Mohanta J. C., Parhi D. R. and Patel S. K. Path Planning Strategy for Autonomous Mobile
Robot Navigation Using Petri-GA Optimisation. ELSEVIER Computers and Electrical
Engineering, 37(6): 1058--1070, 2011.
[94] Kubota N., Morioka T., Kojima F. and Fukuda T. Learning of Mobile Robots Using
Perception-Based Genetic Algorithm. ELSEVIER Measurement, 29(3): 237--248, 2001.
[95] Tuncer A. and Yildirim M. Dynamic Path Planning of Mobile Robots with Improved
Genetic Algorithm. ELSEVIER Computers and Electrical Engineering, 38(6): 1564--1572,
2012.
[96] Ming L., Zailin G. and Shuzi Y. Mobile Robot Fuzzy Control Optimization Using Genetic
Algorithm. ELSEVIER Artificial Intelligence in Engineering, 10(4): 293--298, 1996.
[97] Hu Y., Yang S. X., Xu L. Z. and Meng M. Q. A Knowledge based Genetic Algorithm for
Path Planning in Unstructured Mobile Robot Environments. In IEEE International
Conference on Robotics and Biomimetics (ROBIO), pages 767--772, 2004.
[98] Liu Q., Lu Y. G. and Xie C. X. Fuzzy Obstacle-avoiding Controller of Autonomous Mobile
Robot Optimized by Genetic Algorithm under Multi-obstacles Environment. In IEEE 6th
World Congress on Intelligent Control and Automation, China, pages 3255--3259, 2006.
[99] Li Q., Zhang W., Yin Y., Wang Z. and Liu G. An Improved Genetic Algorithm of
Optimum Path Planning for Mobile Robots. In IEEE 6th International Conference on
Intelligent Systems Design and Applications (ISDA), pages 637--642, 2006.
187
[100] Qu H., Xing K. and Alexander T. An Improved Genetic Algorithm with Co-Evolutionary
Strategy for Global Path Planning of Multiple Mobile Robots. ELSEVIER Neurocomputing,
120: 509--517, 2013.
[101] Algabri M., Mathkour H., Hedjar R., Alsulaiman M. and Al Mutib K. Self-learning Mobile
Robot Navigation in Unknown Environment Using Evolutionary Learning. Journal of
Universal Computer Science, 20(10): 1459--1468, 2014.
[102] Castillo O., Trujillo L. and Melin P. Multiple Objective Genetic Algorithms for Path-
Planning Optimization in Autonomous Mobile Robots. SPRINGER Soft Computing, 11(3):
269--279, 2007.
[103] Arora T., Gigras Y. and Arora V. Robotic Path Planning using Genetic Algorithm in
Dynamic Environment. International Journal of Computer Application, 89(11): 8--12 2014.
[104] Metropolis N., Rosenbluth A. W., Rosenbluth M. N., Teller A. H. and Teller E. Equation of
State Calculations by Fast Computing Machines. The Journal of Chemical Physics, 21(6):
1087--1092, 1953.
[105] Yanar T. A. and Akyurek Z. Fuzzy Model Tuning Using Simulated Annealing. ELSEVIER
Expert Systems with Applications, 38(7): 8159--8169, 2011.
[106] Chang C. C. and Song K. T. Sensor-Based Motion Planning of a Mobile Robot in a

Dynamic Environment. In IEEE International Conference on Industrial Electronics,
Control, and Instrumentation, pages 766--771, 1996.
[107] Martinez-Alfaro H. and Gomez-Garcia S. Mobile Robot Path Planning and Tracking Using
Simulated Annealing and Fuzzy Logic Control. ELSEVIER Expert Systems with
Applications, 15(3): 421--429, 1998.
[108] Zhu Q., Yan Y. and Xing Z. Robot Path Planning Based on Artificial Potential Field
Approach with Simulated Annealing. In IEEE Proceedings of the Sixth International
Conference on Intelligent Systems Design and Applications (ISDA), pages 622--627, 2006.
[109] Precup R. E., Radac M. B., Dragos C. A., Preitl S. and Petriu E. M. Simulated Annealing
Approach to Fuzzy Modeling of Servo Systems. In IEEE International Conference on
Cybernetics, pages 267--272, 2013.
188
[110] Janabi-Sharifi F. and Vinke D. Integration of the Artificial Potential Field Approach with
Simulated Annealing for Robot Path Planning. In IEEE International Symposium on
Intelligent Control, pages 536--541, 1993.
[111] Tavares R. S., Martins T. C. and Tsuzuki M. D. S. G. Simulated Annealing with Adaptive
Neighborhood: A Case Study in Off-Line Robot Path Planning. ELSEVIER Expert Systems
with Applications, 38(4): 2951--2965, 2011.
[112] Liang Y. and Xu L. Global Path Planning for Mobile Robot based Genetic Algorithm and
Modified Simulated Annealing Algorithm. In ACM Proceedings of Summit on Genetic and
Evolutionary Computation, pages 303--308, 2009.
[113] Nakamura E. and Kehtarnavaz N. Optimization of Fuzzy Membership Function Parameters.

In IEEE International Conference on Systems, Man and Cybernetics, pages 1--6, 1995.
[114] Hussein A., Mostafa H., Badrel-din M., Sultan O. and Khamis A. Metaheuristic
Optimization Approach to Mobile Robot Path Planning. In IEEE International Conference
on Engineering and Technology (ICET), pages 1--6, 2012.
[115] Miao H. and Tian Y. C. Robot Path Planning in Dynamic Environments using a Simulated
Annealing Based Approach. In IEEE International Conference on Control, Automation,
Robotics and Vision (ICARCV), pages 1253--1258, 2008.
[116] Zhang Q., Ma J. and Liu Q. Path Planning based Quadtree Representation for Mobile Robot
Using Hybrid-Simulated Annealing and Ant Colony Optimization Algorithm. In IEEE
International World Congress on Intelligent Control and Automation, China, pages 2537--
2542, 2012.
[117] Gao M. and Tian J. Path Planning for Mobile Robot Based on Improved Simulated
Annealing Artificial Neural Network. In IEEE Third International Conference on Natural
Computation, pages 8--12, 2007.
[118] Synodinos A. and Aspragathos N. A. Robot Path Planning of a Mobile Robot using Solid
Modeling Techniques on Potential Fields. In IEEE International Conference on
Mechatronics and Embedded Systems and Applications (MESA), pages 549--553, 2010.
[119] Zhao Y. and Zu W. Real-Time Obstacle Avoidance Method for Mobile Robots Based on a
Modified Particle Swarm Optimization. In IEEE International Joint Conference on
Computational Sciences and Optimization (CSO), pages 269--272, 2009.
189
[120] Castillo O., Martinez-Marroquin R., Melin P., Valdez F. and Soria J. Comparative Study of
Bio-Inspired Algorithms Applied to the Optimization of Type-1 and Type-2 Fuzzy
Controllers for an Autonomous Mobile Robot. ELSEVIER Information sciences, 192: 19--
38, 2012.
[121] Zhang Y., Gong D. W. and Zhang J. H. Robot Path Planning in Uncertain Environment
using Multi-Objective Particle Swarm Optimization. ELSEVIER Neurocomputing, 103:
172--185, 2013.
[122] Zhang Q. and Li S. A Global Path Planning Approach based on Particle Swarm
Optimization for a Mobile Robot. In International Conference on Robotics, Control &
Manufacturing Technology World Scientific and Engineering Academy and Society
(WSEAS), pages 263--267, 2007.
[123] Raja P. and Pugazhenthi S. Path Planning for Mobile Robots in Dynamic Environments
using Particle Swarm Optimization. In IEEE International Conference on Advances in
Recent Technologies in Communication and Computing (ARTCom), pages 401--405, 2009.
[124] Masehian E. and Sedighizadeh D. A Multi-Objective PSO-based Algorithm for Robot Path
Planning. In IEEE International Conference on Industrial Technology (ICIT), pages 465--
470, 2010.
[125] Wong C., Wang H. and Li S. PSO-based Motion Fuzzy Controller Design for Mobile
Robots. International Journal of Fuzzy Systems, 10(1): 284--292, 2008.
[126] Li Q., Tang Y., Wang L., Zhang C. and Yin Y. A Specialized Particle Swarm Optimization
for Global Path Planning of Mobile Robots. In IEEE Third International Workshop on
Advanced Computational Intelligence (IWACI), pages 271--276, 2010.
[127] Huang H. C. FPGA-Based Parallel Metaheuristic PSO Algorithm and its Application to
Global Path Planning for Autonomous Robot Navigation. SPRINGER Journal of Intelligent
& Robotic Systems, 76(3-4): 475--488, 2014.
[128] Chung H. Y., Hou C. C. and Liu S. C. Automatic Navigation of a Wheeled Mobile Robot
using Particle Swarm Optimization and Fuzzy Control. In IEEE International Symposium
on Industrial Electronics (ISIE), pages 1--6, 2013.
190
[129] Shiltagh N. A. and Jalal L. D. Optimal Path Planning for Intelligent Mobile Robot
Navigation using Modified Particle Swarm Optimization. International Journal of
Engineering and Advanced Technology, 2(4): 260--267, 2013.
[130] Chatterjee A. and Matsuno F. A Geese PSO Tuned Fuzzy Supervisor for EKF based
Solutions of Simultaneous Localization and Mapping (SLAM) Problems in Mobile Robots.
ELSEVIER Expert Systems with Applications, 37(8): 5542--5548, 2010.
[131] Juang C. F. and Chang Y. C. Evolutionary-Group-based Particle-Swarm Optimized Fuzzy

Controller with Application to Mobile-Robot Navigation in Unknown Environments. IEEE
Transactions on Fuzzy Systems, 19(2): 379--392, 2011.
[132] Lu L. and Gong D. Robot Path Planning in Unknown Environments using Particle Swarm
Optimization. In IEEE International Conference on Natural Computation (ICNC), pages
422--426, 2008.
[133] Allawi Z. T. and Abdalla T. Y. A PSO-Optimized Type-2 Fuzzy Logic Controller for
Navigation of Multiple Mobile Robots. In IEEE International Conference on Methods and
Models in Automation and Robotics (MMAR), pages 33--39, 2014.
[134] Dorigo M. and Caro G. D. Ant Colony Optimization: A New Meta-Heuristic. In IEEE
International Congress on Evolutionary Computation, pages 1470--1477, 1999.
[135] Guan-Zheng T. A. N., Huan H. and Sloman A. Ant Colony System Algorithm for Real-
Time Globally Optimal Path Planning of Mobile Robots. ELSEVIER Acta Automatica
Sinica, 33(3): 279--285, 2007.
[136] Purian F. K. and Sadeghian E. Mobile Robots Path Planning using Ant Colony
Optimization and Fuzzy Logic Algorithms in Unknown Dynamic Environments. In IEEE
International Conference on Control, Automation, Robotics and Embedded Systems
(CARE), pages 1--6, 2013.
[137] Bi Z., Yimin Y. and Yisan X. Mobile Robot Navigation in Unknown Dynamic
Environment Based on Ant Colony Algorithm. In IEEE WRI Global Congress on
Intelligent Systems, pages 98--102, 2009.
[138] Dong J., Liu B., Peng K. and Yin Y. Robot Obstacle Avoidance based on an Improved Ant
Colony Algorithm. In IEEE WRI Global Congress on Intelligent Systems, pages 103--106,
2009.
191
[139] Ganapathy V., Jie T. T. J. and Parasuraman S. Improved Ant Colony Optimization for
Robot Navigation. In IEEE International Symposium on Mechatronics and its Applications
(ISMA), pages 1--6, 2010.
[140] Fan X., Luo X., Yi S., Yang S. and Zhang H. Optimal Path Planning for Mobile Robots
Based on Intensified Ant Colony Optimization Algorithm. In IEEE International
Conference on Robotics, Intelligent Systems and Signal Processing, pages 131--136, 2003.
[141] Sariff N. B. and Buniyamin N. Comparative Study of Genetic Algorithm and Ant Colony
Optimization Algorithm Performances for Robot Path Planning in Global Static
Environments of Different Complexities. In IEEE International Symposium on
Computational Intelligence in Robotics and Automation (CIRA), pages 132--137, 2009.
[142] Hsu C. C., Hou R. Y. and Wang W. Y. Path Planning for Mobile Robots Based on
Improved Ant Colony Optimization. In IEEE International Conference on Systems, Man,
and Cybernetics (SMC), pages 2777--2782, 2013.
[143] Ganganath N., Cheng C. T. and Tse C. K. An ACO-based Off-line Path Planner for
Nonholonomic Mobile Robots. In IEEE International Symposium on Circuits and Systems
(ISCAS), pages 1038--1041, 2014.
[144] Juang C. F. and Hsu C. H. Reinforcement Ant Optimized Fuzzy Controller for Mobile-
Robot Wall-Following Control. IEEE Transactions on Industrial Electronics, 56(10): 3931-
-3940, 2009.
[145] Hsu C. H. and Juang C. F. Evolutionary Robot Wall-Following Control using Type-2 Fuzzy
Controller with Species-DE-Activated Continuous ACO. IEEE Transactions on Fuzzy
Systems, 21(1): 100--112, 2013.
[146] Juang C. F., Lai M. G. and Zeng W. T. Evolutionary Fuzzy Control and Navigation for
Two Wheeled Robots Cooperatively Carrying an Object in Unknown Environments. IEEE
Transactions on Cybernetics, 45(9): 1731--1743, 2015.
[147] Hsu C. H. and Juang C. F. Multi-Objective Continuous-Ant-Colony-Optimized FC for

Robot Wall-Following Control. IEEE Computational Intelligence Magazine, 8(3): 28--40,
2013.
[148] Chen X., Kong Y., Fang X., and Wu Q. A Fast Two–Stage ACO Algorithm for Robotic
Path Planning. SPRINGER Neural Computing and Applications, 22(2): 313--319, 2013.
192
[149] Hossain M. A., and Ferdousand I. Autonomous Robot Path Planning in Dynamic
Environment using a New Optimization Technique Inspired by Bacterial Foraging
Technique. ELSEVIER Robotics and Autonomous Systems, 64: 137--141, 2015.
[150] Liang X. D., Li L. Y., Wu J. G. and Chen H. N. Mobile Robot Path Planning based on
Adaptive Bacterial Foraging Algorithm. SPRINGER Journal of Central South University,
20: 3391--3400, 2013.
[151] Brand M. and Yu X. H. Autonomous Robot Path Optimization using Firefly Algorithm. In
IEEE International Conference on Machine Learning and Cybernetics (ICMLC), pages
1028--1032. 2013.
[152] Mohajer B., Kiani K., Samiei E. and Sharifi M. A New Online Random Particles
Optimization Algorithm for Mobile Robot Path Planning in Dynamic Environments.
Hindawi Mathematical Problems in Engineering, Article ID 491346: 1--9, 2013.
[153] Bayraktar Z., Komurcu M., Bossard J. A. and Werner D. H. The Wind Driven Optimization
Technique and its Application in Electromagnetics. IEEE Transactions on Antennas and
Propagation, 61(5): 2745--2757, 2013.
[154] Bayraktar Z., Komurcu M., Jiang Z. H., Werner D. H. and Werner P. L. Stub-loaded
Inverted-F Antenna Synthesis via Wind Driven Optimization. In IEEE International
Symposium on Antennas and Propagation (APSURSI), pages 2920--2923, 2011.
[155] Bayraktar Z., Turpin J. P. and Werner D. H. Nature-Inspired Optimization of High-

Impedance Metasurfaces with Ultrasmall Interwoven Unit Cells. IEEE Antennas and
Wireless Propagation Letters, 10: 1563--1566, 2011.
[156] Kuldeep B., Singh V. K., Kumar A. and Singh G. K. Design of Two-Channel Filter Bank
Using Nature Inspired Optimization based Fractional Derivative Constraints. ELSEVIER
ISA Transactions, 54: 101--116, 2015.
[157] Bhandari A. K., Singh V. K., Kumar A. and Singh G. K. Cuckoo Search Algorithm and
Wind Driven Optimization based Study of Satellite Image Segmentation for Multilevel
Thresholding using Kapur’s Entropy. ELSEVIER Expert Systems with Applications, 41(7):
3538--3560, 2014.
[158] Hochsteiner De Vasconcelos Segundo E., Levati Amoroso A., Cocco Mariani V. and Dos
Santos Coelho L. A Wind Driven Approach Using Levy Flights for Global Continuous
193
Optimization. In IEEE International Conference on Artificial Intelligence, Modelling and
Simulation (AIMS), pages 75--80, 2014.
[159] Boulesnane A. and Meshoul S. A New Multi-Region Modified Wind Driven Optimization
Algorithm with Collision Avoidance for Dynamic Environments. In Proceedings of
Advances in Swarm Intelligence, Springer, pages 412—421, 2014.
[160] Mohareri O., Dhaouadi R. and Rad A. B. Indirect Adaptive Tracking Control of a
Nonholonomic Mobile Robot via Neural Networks. Elsevier Neurocomputing, 88: 54--66,
2012.
[161] Hanafi D., Abueejela Y. M. and Zakaria M. F. Wall Follower Autonomous Robot
Development Applying Fuzzy Incremental Controller. Intelligent Control and Automation,
4(1): 18--25, 2013.
[162] Fierro R. and Lewis F. L. Control of a Nonholonomic Mobile Robot Using Neural
Networks. IEEE Transactions on Neural Networks, 9(4): 589--600, 1998.
[163] Cherroun L. and Boumehraz M. Fuzzy Behavior Based Navigation Approach for Mobile
Robot in Unknown Environment. Journal of Electrical Engineering, 13(4): 1--8, 2013.
[164] Chayjan R. A. and Ashari M. E. Modeling Isosteric Heat of Soya Bean for Desorption
Energy Estimation using Neural Network Approach. Chilean Journal of Agricultural
Research, 70(4): 616--625, 2010.
[165] Begum M., Mann G. K. and Gosine R. G. Integrated Fuzzy Logic and Genetic Algorithmic
Approach for Simultaneous Localization and Mapping of Mobile Robots. ELSEVIER
Applied Soft Computing, 8(1): 150--165, 2008.
[166] Contreras-Cruz M. A., Ayala-Ramirez V. and Hernandez-Belmonte U. H. Mobile Robot

Path Planning using Artificial Bee Colony and Evolutionary Programming. ELSEVIER
Applied Soft Computing 30: 319--328, 2015.
[167] Mohanty P. K. and Parhi D. R. Optimal Path Planning for a Mobile Robot using Cuckoo
Search Algorithm. TAYLOR & FRANCIS Journal of Experimental & Theoretical Artificial
Intelligence, 28(1-2): 35--52, 2014.
[168] Vakil-Baghmisheh M. T. and Navarbaf A. A Modified Very Fast Simulated Annealing

Algorithm. In IEEE International Symposium on Telecommunications, pages 61--66, 2008.
194
[169] Jianguo W., Yilong Z. and Linlin X. Adaptive Genetic Algorithm Enhancements for Path
Planning of Mobile Robots. In IEEE International Conference on Measuring Technology
and Mechatronics Automation, pages 416--419, 2010.
[170] Liu C., Gao Z. and Zhao W. A New Path Planning Method Based on Firefly Algorithm. In
IEEE Fifth International Joint Conference on Computational Sciences and Optimization
(CSO), pages 775--778, 2012.
[171] Motlagh O. R. E., Hong T. S. and Ismail N. Development of a New Minimum Avoidance
System for a Behavior-Based Mobile Robot. ELSEVIER Fuzzy Sets and Systems, 160(13):
1929--1946, 2009.
[172] Mohanty P. K. and Parhi D. R. Navigation of Autonomous Mobile Robot Using Adaptive
Network Based Fuzzy Inference System. SPRINGER Journal of Mechanical Science and
Technology, 28(7): 2861--2868, 2014.
[173] Deepak B. B. V. L., Parhi D. R. and Raju B. M. V. A. Advance Particle Swarm

Optimization-Based Navigational Controller for Mobile Robot. SPRINGER Arabian
Journal for Science and Engineering, 39(8): 6477--6487, 2014.
[174] Tahmasebi P. and Hezarkhani A. A Hybrid Neural Networks-Fuzzy Logic-Genetic

Algorithm for Grade Estimation. ELSEVIER Computers & Geosciences, 42: 18--27, 2014.
195
Dissemination
Journal Articles
1. Anish Pandey, and Dayal R. Parhi, “Multiple Mobile Robots Navigation and Obstacle
Avoidance Using Minimum Rule Based ANFIS Network Controller in the Cluttered
Environment”, SOJ Robotics Automation, 1(1): 1--11, 2016.
2. Anish Pandey and Dayal R. Parhi “New Algorithm for Behaviour-Based Mobile Robot
Navigation in Cluttered Environment Using Neural Network Architecture”, World
Journal of Engineering, 13(2): 129--141, 2016.
3. Anish Pandey and Dayal R. Parhi “Autonomous Mobile Robot Navigation in Cluttered
Environment using Hybrid Takagi-Sugeno Fuzzy Model and Simulated Annealing
Algorithm Controller”, World Journal of Engineering, Accepted, Paper ID-
wje20160115003.
4. Dayal Ramakrushna Parhi, and Anish Pandey, “A Review of Mobile Robot Localization
and Navigation”, International Science Press IJAAIES, 4(2): 103--109, 2012.
5. Dayal R. Parhi, Anish Pandey, and Rakesh Kumar Sonkar, “A Review of Mobile Robot
Navigation Using a Fuzzy Logic”, International Science Press IJAICR, 5(2): 103--108,
2013.
Conference Presentations
1. Anish Pandey, Rakesh Kumar Sonkar, Krishna Kant Pandey, and D. R. Parhi, “Path
Planning Navigation of Mobile Robot with Obstacles Avoidance Using Fuzzy Logic
Controller”, In IEEE 8th International Conference on Intelligent System and Control
(ISCO), Coimbatore, pages 36--41, 2014.
2. Anish Pandey, and Dayal Ramakrushna Parhi, “MATLAB Simulation for Mobile Robot
Navigation with Hurdles in Cluttered Environment Using Minimum Rule Based Fuzzy
Logic Controller”, In ELSEVIER Procedia Technology, 2014, Vol. 14, pages 28--34.
3. Anish Pandey, Rakesh Kumar Sonkar, P. K. Mohanty, and D. R. Parhi, “Navigation of

Mobile Robot with Obstacles Avoidance Using Fuzzy Logic”, In Proceedings of
196
International Conference on Advances in Mechanical Engineering, COEP, Pune,
Maharashtra, Paper ID ICAME-S1/P1, May 29-31, 2013.
4. Prases K. Mohanty, Dayal R. Parhi, Alok K. Jha, and Anish Pandey, “Path Planning of
an Autonomous Mobile Robot using Adaptive Network Based Fuzzy Controller”, In
IEEE 3rd International Conference on Advance Computing Conference (IACC), pages
651--656, 2013.
5. Anish Pandey, Saroj Kumar, Krishna Kant Pandey, Dayal R. Parhi, “Mobile robot
navigation in unknown static environments using ANFIS controller”, In: ELSEVIER
Perspectives in Science, dx.doi.org/10.1016/j.pisc.2016.04.094, 2016.
Communicated Articles
1. Anish Pandey and Dayal R. Parhi, “Mobile Robot Optimum Path Planning in Cluttered
Environment using Wind Driven Optimization Algorithm”, SPRINGER Soft Computing,
Article No. SOCO-D-15-00890.
2. Dayal R. Parhi and Anish Pandey, “Optimum Path Planning of Mobile Robot in
Unknown Static and Dynamic Environments using Fuzzy-Wind Driven Optimization
Algorithm”, SPRINGER Intelligent Service Robotics, Article No. JIST-D-16-00018.
Patent
1. Automated Four-Wheeled Intelligent Stair Climbing Mobile Robot (Application No.:

201631009662) Inventors: 1. Dr. Dayal R. Parhi, 2. Mr. Anish Pandey (filed).
197
Vitae
Mr. Anish Pandey was born in 1988 at Bilaspur district of Chhattisgarh (C.G.). He has
completed his class X and class XII from G. T. B. Higher Secondary School, Bilaspur
(C.G.) in 2003 and 2005, respectively. He completed his Bachelor of Engineering in
Industrial and Production Engineering from Guru Ghasidas Central University, Bilaspur,
in 2009 and his Master of Engineering in Mechanical Engineering (specialization in
machine design) from Chhattisgarh Swami Vivekananda Technical University, Bhilai, in
2011. After the completion of his Post-Graduation he has joined Ph.D. program at
Department of Mechanical Engineering, National Institute of Technology, Rourkela in
July 2012 and submitted the Ph.D. thesis in April 2016. His research interest includes
Mobile Robot Navigation, Humanoid Robots, and Autonomous Aerial Robots.
198
Appendix-A
Figure A1: Arduino microcontroller based experimental differential drive mobile robot.
199
Table A1: Specifications of the experimental mobile robot
Name Specifications
Microcontroller Arduino MEGA 2560 (ATmega2560).
Flash Memory 256 KB (ATmega2560).
Operating Voltage 5V.
SRAM 8 KB (ATmega2560).
Input Voltage 7–12V (Recommended).
Input Voltage (Limits) 6–20V.
Digital Input/Output
54 (of Which 15 Can be Used as PWM Outputs).
Pins
Analog Input Pins 16.
Motors 2 DC, 30RPM Centre Shaft Economy Series DC Motor.
Motors Driver L298, Up to 46V, 2A Dual DC Two Motor Drivers.
Motor Speed Max: 30RPM, Min: 12RPM.
Wheel Diameter: 106mm, Wheel Thickness: 44mm, Hole

Wheel
Diameter: 8mm.
One Sharp Infrared Range Sensor (GP2Y0A02YK0F)
Distance Measuring Range: 20cm to 150cm.
Sensors
Two Ultrasonic Range Finder Sensor (HC SR-04) Distance
Measuring Range: 2cm to 400cm.
Bread Board Small Size Bread Board.
Communication USB connection Serial Port.
Principal Dimensions Height: 30cm, Length: 62cm, Width: 49cm.
Weight Approx. 4.936kg.
Payload Approx. 500g.
Two Rechargeable Lithium Polymer 3 Cell, 11.1V,

Power
2000mAh, 20C Battery
200
201
View publication stats

Thesis AnishPandey 512ME119

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thesis AnishPandey 512ME119

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Mobile Robot Navigation in Static and Dynamic Environments using Various

Thesis · July 2016

Dr. Anish Pandey

Wheeled cleaning robot View project

The user has requested enhancement of the downloaded file.

Department of Mechanical Engineering

Dissertation submitted to the

National Institute of Technology Rourkela

in partial fulfillment of the requirements

(Roll Number: 512ME119)

under the supervision of

Prof. Dayal R. Parhi

Department of Mechanical Engineering

July 18, 2016

Name: Anish Pandey

Title of Dissertation: Mobile Robot Navigation in Static and Dynamic Environments

Bidyadhar Subudhi Tarapada Roy

Sarat Kumar Das Santosha Kumar Dwivedy

Dr. Dayal R. Parhi

July 18, 2016

July 18, 2016

July 18, 2016 Anish Pandey

Keywords: Intelligent Mobile Robot; Navigation; Hybrid Fuzzy; Cascade Neuro-Fuzzy;

Supervisor’s Certificate iii

List of Figures xiii

Symbols & Abbreviations xxiii

1.1 Background and Motivations 1

1.2 Aims and Objectives of the Proposed Research Work 2

1.3 Methodologies Applied for Proposed Research Work 2

1.4 Novelty of the Proposed Research Work 3

1.5 Outline of the dissertation 4

2 Literature Review 5-24

2.2 Kinematic and Dynamic Analysis of the Wheeled Mobile Robot 7

2.3 Various Soft Computing Techniques used for Mobile Robot 8

2.3.1 Fuzzy Logic Technique for Mobile Robot Navigation 8

2.3.2 Neural Network Technique for Mobile Robot Navigation 12

2.3.3 Neuro-Fuzzy Technique for Mobile Robot Navigation 14

2.3.5 Simulated Annealing Algorithm for Mobile Robot 18

2.3.6 Particle Swarm Optimization Algorithm for Mobile Robot 20

2.3.7 Ant Colony Optimization Algorithm and Other 21

2.3.8 Wind Driven Optimization Algorithm 23

3 Kinematic and Dynamic Analysis of the Nonholonomic Differential 25-33

3.2 Kinematic Model of the Nonholonomic Differential Drive Two- 25

3.3 Dynamic Model of the Nonholonomic Differential Drive Two- 32

4 Intelligent Navigation of a Mobile Robot in Static and Dynamic 34-59

4.2 Hybrid Fuzzy (H-Fuzzy) Architecture 35

4.2.1 Takagi-Sugeno Type Fuzzy Logic Architecture (TFa) for 36

4.2.2 Mamdani-Type Fuzzy Logic Architecture (MFa) for Obstacle 40

4.3 Simulation Studies 44

4.4 Comparison with Previous Works 50

4.5.1 Arduino Microcontroller based Wheeled Mobile Robot 53

5 Intelligent Navigation Control of a Mobile Robot in Unknown 60-84

5.2 Cascade Neuro-Fuzzy (CN-Fuzzy) Architecture 61

5.2.1 Cascade Neural Network for Goal Reaching 62

5.2.2 Fuzzy logic architecture (FLA) for obstacle avoidance 65

5.3 Computer Simulation Results 69

5.4 Comparison with Previous Works 73

5.4.1 First Comparison with Previous Works 73

5.4.2 Second Comparison with Previous Works 75

5.5 Experimental Results 77

5.5.1 Experimental Mobile Robot Description 77

6 Mobile Robot Navigation in Different Environments using Takagi- 85-113

6.2 Design of Sugeno-Type Fuzzy Logic Controller 87